Semantic Memory Overview

Semantic Memory

• Semantic knowledge allows humans to predict, understand, and respond to new experiences
• It often arises from common elements of multiple experiences
• Scientists use formal methods to organize knowledge about the world
• Semantic memory needs to support fact retrieval and reasoning about relationships between facts
• It also needs to support inferences about new experiences
• Semantic networks organize knowledge hierarchically, with distances representing relatedness
• Retrieval happens through spreading activation in the network
• Semantic fluency tests measure category retrieval speed and order, reflecting semantic relatedness
• Animal concepts are organized in semantic memory based on size and distance (how often people generated the concept) and how often two concepts are generated one after another
• Priming in semantic memory happens when exposure to a stimulus influences a subsequent stimulus response, even without conscious awareness
• Priming is done by presenting a prime stimulus during an unrelated task. The activation spreads to other related concepts, influencing behavior
• Mental representations of visual objects (like faces) are complex layers of features organized by increasingly complex abstractions based on visual input
• Words are composed of a fixed set of features (letters)

Recognizing Words

• Comprehending text involves recognizing words and retrieving their meaning from semantic memory
• Interactive activation theory describes how words are represented as combinations of letters, and vice-versa. These combinations also use line segments (features) as components that activate and spread within the network. Activation spreads in both directions, influencing word and letter recognition.
• Words are recognized faster within an actual sentence versus a scrambled sentence or in an unfamiliar context

Conceptual Knowledge

• Conceptual knowledge is organized in hierarchical associative networks.
• Retrieval of semantic memory happens through spreading activation.
• Lower-level features and higher-level concepts interact in representations.
• Word superiority effect happens when recognizing words within a sentence increases recognition speed.
• Activation of words can be very useful in recognizing individual parts of the word.
• Semantic memory is de-contextualized, organized by meaning, and generalizable.

Category Learning

• Category membership is graded (some items are better examples than others).
• Family resemblance property: category members share overlapping features.
• Prototype theory describes a concept as the central tendency of common features.
• Prototypes are used to evaluate new objects and determine category membership
• Similarity to the prototype affects how quickly an item is categorized
• Categorization happens by comparing new stimuli to the stored prototype
• Features can be physical or non-physical
• Ambiguous or complex objects may be equidistant to multiple categories, and often harder to classify
• Solso and McCarthy’s experiment showed that people are quicker to recognize highly similar faces compared to faces with little resemblance to a prototype
• Categorizing involves comparing a new object to previously learned features
• Categorization is not strictly all-or-nothing--some items are considered better examples than others.
• Prototypes are the most common, ideal examples or features
• Recognition of unusual features are less accurate than the other features of the prototype
• Prototype theory is a concept theory that combines the most common features to recognize a category

Generalization

• Generalization occurs when past learnings are applied to new situations based on perceived similarities of past experiences.
• Generalization uses perceived similarities to predict consequences based on past experiences.
• Discrimination learning distinguishes how animals/people learn to react to different stimuli.
• Discrimination learning is important to learning how natural categories based on different outcomes, members are associated with different outcomes.
• Through experience, one can deduce a relationship between features and different outcomes.

Discrimination Learning

• Discrimination learning is a form of learning in which individuals learn to respond differently to different stimuli.
• It's critical for classifying natural categories in meaningful ways.
• Through experience, categories are associated with different outcomes (i.e., pigeons learned to discriminate between two tones based on different reward outcomes.)

This quiz explores the intricacies of semantic memory, including how semantic knowledge aids in predicting and responding to experiences. It covers concepts like semantic networks, retrieval processes, and the role of priming in memory. Test your understanding of how we organize and retrieve knowledge about the world.