These notes represent a detailed interpretation of the professor’s lecture. GradeBuddy is best used as a supplement to your own notes, not as a substitute. Lecture 15 Outline of Last Lecture I. Knowledge II. Categorization III. Prototype Approach IV. Exemplar Approach V. Hierarchical Organization VI. Semantic Networks Outline of Current Lecture I. Semantic Networks II. Properties of Good Psychological Theories III. The Connectionist Approach IV. Categories in the Brain V. Human Development and Categorization Current Lecture I. Semantic Networks a. Concepts are arranged in networks that represent the way concepts are organized in the mind b. Collins and Quillan (1969) i. Node = category/concept ii. Concepts are linked iii. Model for how concepts and properties are associated in the mind c. Critiques of Collins and Quillan i. Cannot explain typicality effects ii. Some evidence against cognitive economy in sentence-verification tasks d. Spreading Activation i. Activation is the arousal level of a node ii. When a node is activated, activity spreads out along all connected links iii. Concepts that receive activation are primed and more easily accessed from memory iv. Collins and Loftus (1975) modifications PSYC 310 1st Edition1. Shorter links to connect closely related concepts 2. Larger links for less closely related concepts 3. No hierarchical structure, based on person’s experience v. Cognitive Economy: Shared properties are only stored at higher level nodes vi. Inheritance: 1. Lower level items share properties of higher level items 2. Cognition must travel up the hierarchy to retrieve info from stored at higher nodes vii. Lexical Decision Task 1. Participants read stimuli and are asked to say as quickly as possible whether the item is a word or not II. Properties of Good Psychological Theories a. Explanatory power – theory can explain why an experimental result occurred b. Predictive power – theory can predict future result given info on the controlling variables c. Generation of Experiments – many researchers conduct studies to evaluate and improve the theory d. Falsifiability – must be possible to specify what it would take to disprove the theory empirically i. “If this happens, then the theory is right. If that happens, then it’s wrong.” III. The Connectionist Approach a. Neuron-like units (sometimes called nodes) i. Input units: activated by stimulation from environment ii. Hidden units: received input from input units iii. Output units: receive input from hidden units iv. Each connection between units is assigned its own “weight” that determines how strongly info is passed between them 1. Positive weight causes excitation; negative causes inhabitation v. How learning occurs 1. Network responds to stimulus 2. Provided with correct response a. Error signal: Difference between actual activity of each output unit and the correct activity b. Modifies weights slightly to better match correct responsec. Next time the network gets that stimulus, it is more likely to answer it correctly 3. Slow learning process that creates a network capable of handling a wide range of inputs 4. Learning can be generalized 5. Graceful degradation: disruption of performance occurs gradually as parts of the system are damaged IV. Categories in the Brain a. Different areas of the brain may be specialized to process information about different categories i. Double Dissociation for categories “living things” and “nonliving things” b. Categories are represented by distributed activity c. Category specific neurons V. Human Development and Categorization a. Using familiarization/preference procedure. Quinn & Johnson found: i. Infants as young as 2 months can categorize broad categories (mammals vs furniture) ii. By 3-4 months basic level categories partially developed (dogs vs. cats) iii. 6-7 months: more specific categories (tabby vs Siamese cate) with