The Constructionist-Integration Model of Reading

The Constructionist-Integration (Kintsch, 1988; 2004) Model of Reading comprehension (CI) sounds like it may be a construction crew out building libraries or bookstores, but it is really a model of how working memory constructs meaning by assembling propositions (sentences or meaning units like phrases).  This model is from the constructivist school of thought, which espouses the idea that we create meaning from the world around us rather than discover a meaning that was intended from the beginning.  In other words, a reader brings his or her background knowledge to the text, and in that manner, he or she will have a unique reading experience no matter the book.  If a sentence says "dog," one child might envision a poodle while another envisons a pit bull.  Therefore, each reader interacts with and experiences the text differently.

In order to understand this model, it is necessary to understand some key terminology.

• textbase or microstructure--the current proposition and 2-3 surrounding sentences
• macrostructure--the gist of the text as a whole
• situation model--a mental model of the text as it is being read; the reader visualizes the text using his/her background knowledge and supplies details where the details are scant.  For example, a reader might think of the smell of an apple pie in the story even if the smell hasn't been discussed.  Because situation models are unique to the individual, they vary from reader to reader.
• local coherence--thought to be the main goal of the comprehension process in a minimalist model; a reader will make a minimal amount of inferences to connect seemingly unrelated sentences, for example. Local coherence is maintained with propositions that are in close proximity to one another.
• global coherence--  thought to be the process used in a constructionist model; the reader is continually making inferences as he/she reads in an effort to make sense of the text of the whole and even retrieves previously suppressed material as it becomes relevant.

Context plays a pivotal role in the construction of propositions.  According to Kintsch (1983) and his macro rules that govern the encoding of propositions, context determines what meanings are quickly suppressed, what propositions are deleted from the microstructure, if a series of propositions can be reduced to one general statement, or if one proposition activates another (i.e. inferring "school" from "grades").  Because English has so many polysemous words, context is key to processing the meaning of a text, as Adams (1990) inclusion of  a context processor in her reading model indicates.

Kintsch

For further reading:
Inference Generation and Cohesion in the Construction of Situation Models:  Some Connections with Computational Linguistics  (Graesser et al., 2007)