Computational Model for Narrative and Event Comprehensions
Published:
Living in Our Modern Industrialized Society
In our modern industrialized society, we engage in a plethora of dynamic activities, presented in various formats, from visual and auditory experiences to text-based reading comprehensions. Whether we are reading a descriptive paragraph about a little boy named Jamie starting his day or watching a movie depicting an actor’s morning routine, one may wonder if these activities require similar cognitive processes and architectures. While some lower-level processes, such as orthographic processes and decoding, are domain-specific, other cognitive processes extend beyond the medium of text (Magliano, Joseph P et al., 2007).
Article Aims
This article aims to introduce various comprehension models for narrative comprehensions and offer alternative representations for event cognition. We will argue that the success of remapping theories of reading comprehension to dynamic cognition may depend on the goals of reading comprehension and the content of the texts being comprehended.
Reading Comprehension
In the realm of reading comprehension, two types of inferences play a critical role: bridging inferences and associative learning or elaborations. Bridging inferences are crucial for understanding the causal structures within narrative stories, while elaborations are more critical for grasping concepts and instances.
The Construction-Integration Model
Historically, the first model proposed for reading narrative texts is the Construction-Integration model (Kintsch, 1988). For reading comprehension, the activation of information in the text and related knowledge is essential. Similarly, when observers watch a movie, they experience a slowdown in processing as they attempt to construct a situation model. Once a mental representation is established, it needs to be modified to represent the comprehended text fully. This process results in greater activation for concepts linked to other concepts, creating a unique activation pattern. While the CI model is best suited for explaining learning from text, it does not address mechanisms and features involved in understanding passages that depict relatively familiar scenarios found in narrative text.
Structure-Building Model
The Structure-Building model, proposed by Gernsbacher(1990, 1997), added an additional mechanism compared to the Construction-Integration model. In addition to laying a foundation and mapping information, it also involves shifting to new structures due to incongruencies. According to this model, shifting due to incoming new information is not necessarily optimal, as suppression mechanisms are critical for readers to engage with the current situational models and effectively ignore incongruities. This notion may also apply to processing dynamic activities, where representing different variations of actions as a single action is important. However, this model does not address when shifting based on new information becomes more optimal during online comprehension.
The Resonance Model
The Resonance Model (Myers & O’Brien, 1998) aims to explain how information relatively distant from the focal sentence can be used for comprehension. It suggests that information in working memory serves as indexes for information stored in long-term memory. For comprehending activities that last several hours, this ability might be critical as well.
The Event-Indexing Model
The Event-Indexing Model (Zwaan & Radvansky, 1998) assumes that the cognitive system is attuned to perceive dynamic events and monitors coherence along five dimensions: time, space, causality, motivation, and agents. It involves three stages: the current model, the integrated model, and the complete model. The Event-Indexing Model closely resembles our experiences in a dynamic world, but it is primarily a model for reading comprehension with clear beginnings and endings.
Summary
In summary, the goal of comprehending narrative texts appears to differ from understanding events, as theories of event segmentation focus on breaking down continuous streams of experiences into episodic segments. In contrast, comprehension of narrative inferences aims to bridge inferences and map current stimuli into existing mental representations. However, a flexible yet comprehensive comprehension model should be able to address both perspectives, bridging the gap between literature traditions and cognitive science.
References
- Gernsbacher, M. A. (1997). Two decades of structure building. Discourse Processes, 23(3), 265–304. DOI
- Gernsbacher, M. A., Varner, K. R., & Faust, M. E. (1990). Investigating Differences in General Comprehension Skill.
- Kintsch, W. (1988). The Role of Knowledge in Discourse Comprehension: A Construction-Integration Model.
- Magliano, Joseph P, Radvansky, Gabriel A, & Copeland, David E. (2007). Beyond language comprehension: Situation models as a form of autobiographical memory. Higher Level Language Processes in the Brain: Inference and Comprehension Processes, 379–391.
- Myers, J. L., & O’Brien, E. J. (1998). Accessing the discourse representation during reading. Discourse Processes, 26(2–3), 131–157. https://doi.org/10.1080/01638539809545042
- Zwaan, R. A., & Radvansky, G. A. (1998). Situation Models in Language Comprehension and Memory.