Fathom: The Source for Online Learning  
 
Help About Us Course Directory
Browse Fathom


 
 
 
Types of Mental Representation
From: Cambridge University Press | By: Walter Kintsch

EDITOR'S INTRODUCTION | Human mental representation works on a variety of levels, from the ability simply to recall something from our environment to the most abstract thought. Yet no definitive account of it exists. In this extract from his book Comprehension, psychologist Walter Kintsch attempts to categorise the diversity of such representations and to assess their interrelations.


A definitive account of mental representations does not yet exist.
ne can think of mental representations as forming a hierarchy of abstractness and increasing independence from the environment. The most basic forms of mental representations are procedural and perceptual representations that are tied directly to the environment. Episodic memory representations are at the next higher level. Intentional but nonverbal representations, including forms of imagery, make up the next layer of mental representations. The final two layers achieve the most independence from the environment and are both verbal: the verbal narrative level and the verbal abstract level. The hierarchy is not one of complexity, however: There are highly complex perceptual representations and very simple abstract representations. The defining feature of the hierarchy is that it changes from direct representations of the environment to ever more indirect, flexible ones that permit more and more arbitrary, unconstrained computations. The layers of the hierarchy are ordered in terms of their appearance on the mental stage, both phylogenetically and ontogenetically. As new forms of mental representations arise, the older forms do not disappear but remain embedded within the newer layer. Thus, the picture is one of gradual unfolding of the full capacity of the human mind. It is a picture the details of which are still quite vague, and as these details become better understood, some of the picture's features may have to be corrected. But the implications of the framework suggested here for cognitive theory remain relevant, even if the framework is altered in detail.

From bottom to top

As one moves from the bottom to the top layers of mental representation, the general character of the representations changes in the following ways.
  • Most significantly, the degree of environmental control weakens; stories and mathematical theories liberate man from the web of environmental dependencies.
  • Representations change from sensorimotor and analog in the lower layers to symbolic and arbitrary in the upper layers. In between, an image or a gesture may be used in a symbolic but nonarbitrary way.
  • At the same time, the degree of consciousness increases. Similarly, the degree of intentionality increases. Thus, we have at the one extreme procedural representations that are acquired through incidental, unconscious learning, and at the other extreme formal school learning that requires intentional and conscious acts.


Direct procedural and perceptual representations. These forms of representations involve largely innate systems. Many different types of affordances, abilities, and actions exist, as well as different biological mechanisms. These forms of representation can be modified through experience. Responses to environmental affordances can be modified by perceptual or procedural learning. Such learning is tightly coupled to the environment (e.g., how to tie a shoelace). Repetition and reinforcement determine the learning process within constraints imposed by the abilities of the organism and the affordances of the environment. This capacity is shared by all animals.


Episodic representations. Cognition at this level is based on episodic memory representations, that is, generalized event representations of experience that are created to guide action and anticipate changes in the environment. Event memory, unlike procedural memory, is accessible to recall and reflection, because it is a form of declarative memory. Out of event memory emerges the recollection of particular experiences, involving a certain level of consciousness and self-awareness. It permits the analysis and breakdown of perceptual events. Event memory is shared with higher animals. What is remembered and represented in memory are concrete events and scriptlike sequences of events. Learning occurs through experience and is incidental, unintentional, and goal-directed.


Organisms, such as apes, that must rely on event memory because they do not have higher forms of cognition, use signs and have a repertoire of social skills that depends on a rich episodic memory, but their actions are directly linked to the environment. Thus, cognition at this level is analytic and reflective but still environmentally bound. Human event memory is different from the "pure" event memory of apes, because humans can code their memories linguistically. Human event memory is embedded in other forms of representations. As a consequence, human event memory is tied in with higher cognition, that is, with linguistic and symbolic thought.


Nonverbal, imagery, and action representations. These representations are sensorimotor in character but are used intentionally, often but not necessarily for the purpose of communication, especially the communication of emotions (e.g., body language). An established social community is a prerequisite for representations of this type.


Narrative oral representations. These representations are one of two types of linguistic representation. They are verbal but not abstract. Their structure is linear, and information processing at this level is analytic and rule-governed, as in semantic memory, prepositional memory, discourse comprehension, analytic thought, induction, and verification.


Much of what we know and what we learn is in the form of stories--for example, our cultural and historical knowledge. Stories are narrative mental models that allow us to learn about the world. The world becomes more comprehensible to us when we are able to tell a coherent story about it. There is again a social component to narrative learning because stories are told by someone to someone (including to oneself). Socially elaborated and sanctioned stories are the cognitive structures that hold a culture together.


Abstract representations. Abstract representations are required for categories, logical thought, formal argument, deduction, quantification, and formal measurement. Abstract symbols are dependent on visuographic invention: written language, maps, calendars, clocks, artistic and scientific graphing, and other forms of external memory storage. Knowledge is primarily stored in the world, not in the individual brain. Biological memory carries around the code for the use of external memory, whereas the specifics are found in external symbolic storage systems. Pedagogy has always been directed at this abstract level. This is where learning problems arise and where most special instructional efforts are needed. This is what school learning is about.

Aspects of meaning

To illustrate these distinctions with an example, consider the way people assign meaning to one object by transferring to it aspects of meaning from another. At the nonverbal level, an example is children's pretense play: A banana becomes a "telephone" by the child's talking into it, as adults do with a telephone. An example at the narrative linguistic level is metaphor: One concept assumes certain semantic features from another. For example, sermons assume some properties of sleeping pills. Analogy is the corresponding example at the abstract linguistic level: Objects are no longer directly compared, but the meaning transfer depends on an abstract categorical system (the analogy between electricity and water flowing through pipes requires a precise mapping of categories).


The neat classification presented here is complicated by the fact that, in the adult human, lower forms of cognition are encapsulated by the higher forms. The types of cognition we have discussed are found in pure form in the phylogenetic development of man or in the ontogenetic development of cognition, but not in the adult human mind. Thus, people have incidental event memory, as a dog might have, but they also have explicit, language-coded event memory that is uniquely human. Even procedural memory may be language-coded, though not very well, as is testified by the inefficiency of verbally describing a tennis stroke or a skiing turn. Action representations, too, are coded with language. Thus, a modern apprenticeship situation is not purely imitational but supplemented by oral language and symbolic thought, as, for example, a graduate student doing work in some laboratory learns skills.


Conversely, the higher layers of representation that characterize the modern adult human mind do not exist in a symbolic vacuum but are based on the sensorimotor substratum that humans share with other forms of animal life. Hence, language cannot be understood only as an arbitrary disembodied linguistic system and cannot be separated from its nonlinguistic substratum.