09 Apr Abstract Contexts and Fuzzy Reasoning
We do not yet know how we remember things, nor do we know how we use remembered things in reasoning. The amazing feedback loops of afferent and efferent fibers between different layers of the cortex give us some amazing clues (Hawkins 2004). Today’s discussion of abstract contexts and fuzzy reasoning is intended as a bridge between this section on cognition and the upcoming section on cybernetic modeling. Some experts describe the organization of conceptual memory in terms of schemata. Frederick C. Bartlett described schemata as collective representations of reactions or experiences with some commonality. This may be topical commonality, experiential or historical commonality, or associative commonality. The schematic memory model can be described using some of the recurring terms of this study. For example, topical commonalities could be described as contexts or domains.
A schema would constitute the organization of knowledge about a single domain or a specific type of past experience: a cluster of associated concepts all bound together through part-whole relations, causal relations, or some other set of shared attributes (Sowa 1984). Domains may be physical or abstract, and within domains, the range of associated concepts may be both physical and abstract. For example, in the abstract domain of interpersonal communication, sounds are physical, generated by physical organs, conveyed through the air as physical waves, and received by physical mechanisms in the inner ear that physically respond to the vibrations. The intent of the speaker, however, is abstract. In the physical domain of building construction, budgets and schedules are abstract, while most other concepts represent physical things.
|Understanding Context Cross-Reference|
|Click on these Links to other posts and glossary/bibliography references|
|Prior Post||Next Post|
|Symmetrical Logic and Lineage||Are you who you say you are?|
|conceptual schemata||Thorndyke 1979 Hawkins 2004|
|abstract knowledge domain||Schank 1976 Bartlett 1932|
|experiences association||Sowa 1984 Carbon 2012|
Schemata (plural of schema) have alternatively been described as defining some “more complex or frequently encountered concept” (Thorndyke, 1979, p. 82). I have illustrated a schema for a domain as a grouping of associated concepts, each concept binding objects in the domain together with specific relations. In this illustration, the concepts in a domain are disjoint. In more advanced schemata, the concepts are linked. In the brain, nothing is disjoint. When we think about things by accessing memory of learned concepts, the context of the input, whether visual, tactile or linguistic, governs or constrains the natural flow of impulses throughout the brain. This does not suggest that a schema or context grouping occupies some contiguous region in the brain. The grouping would probably be mediated by the strength or tightness of the connections rather than by the co-location of the clusters of neurons representing conceptual components of the schema.
The question of how we determine context is a question separate from the topic of this post, but consider the contextual agility of comedians or punsters who leverage an ability to evoke multiple or unusual contexts to cause us to jump out of our normal contextual mold and find an exotic connection.
Would you trust a computer system that can complete your sentences for you? Script theory can help explain constructive elaboration, a common phenomenon among husbands and wives and others who share experiences. A script is part of a schema in which stereotyped knowledge of predictable events or responses is stored. For example, whenever my Father in-law, Grant drives past a cemetery, he says, “There’s a grave situation.” When recalling a specific episode, I may recall Grant saying “There’s a grave situation” on that specific occasion when, in fact, it was I who said it preemptively.
Constructive elaboration can explain why any number of witnesses to a single event are likely to have some differences in their retelling of the events. Roger Schank describes a script as a “causal chain of conceptualizations that have been known to occur in that order many times before” (1976, p. 180). If a script exists as part of a schema, its memory can be triggered by an image, sound, smell, or a word or phrase in context. Scripts can even be triggered by thought processes without any perceptual input in context. For example, in the course of thinking about something unrelated, a script may be triggered by overlapping schema.
Schema and/or scripts for common contexts or domains may be shared across groups of people who share common types of experiences. Their research was among students. Because the individual experiences of the students were, in large part, determined by the collective nature of their cultural background (which is, in itself, contextual), the similarities in these scripts were distinct and measurable.
The hierarchical organization of information in the physical universe resembles the organization of knowledge in our brains: all the information is associated in hierarchical and causal connections. In the universe, the connections are governed by physical laws. In the brain, the connections are learned and somehow stored in the network.
Functionally, our brains use logic to figure things out. This logic comes in various forms, including binary, multi-valued, heuristic, and threshold logic. We also reason about the structure of the universe and the information we encounter, and we seek symmetries.
Since the brain is so slow compared to computers, there has to be something that simplifies the reasoning process so we can make decisions quickly enough to avoid danger. From the time we wake up and begin our day, our brains continually keep track of where we are and what we are doing. This context keeps applicable areas of our brain warmed up so it is easier for new information to generate hot spots. This phenomenon results in expectations. When we are driving on a road, we expect signs that tell us what to do and what not to do. Because of these expectations, we are prepared to act more quickly, and we can recognize the meaning of signs more quickly.
The understanding of brain functions and cognitive processes provides a foundation for the information in the next sections in which I will propose a simple, but elegant model for understanding intent, and using this understanding as a basis for more accurate answers to complex questions, intelligent dialog between humans and devices, and greater accuracy in translating texts from one language to many others.
Here are some great references in this subject area:
|Click below to look in each Understanding Context section|
|4||Perception and Cognition||5||Fuzzy Logic||6||Language and Dialog||7||Cybernetic Models|
|8||Apps and Processes||9||The End of Code||Glossary||Bibliography|