Cognitive Psychology Lecture Notes PDF

# Cognitive Psychology Lecture Five - Concepts and Semantic Memory ## You should understand * What are concepts? * The prototype approach * The exemplar approach * The knowledge-based approach * Using concepts * Organisation of concepts * Schemas and use of concepts in our social behaviour ## Lo...

# Cognitive Psychology Lecture Five - Concepts and Semantic Memory ## You should understand * What are concepts? * The prototype approach * The exemplar approach * The knowledge-based approach * Using concepts * Organisation of concepts * Schemas and use of concepts in our social behaviour ## Long-term memory ### Long-term memory * **Explicit (declarative)** * **Episodic (biographical events)** - e.g. remembering that the word ‘elephant’ was presented in a studied list of words or remembering getting your A-level results (Endel Tulving, 1972) * **Semantic (words, ideas, concepts)** - knowing that elephants live in Africa or knowing your A-level results * **Implicit (non-declarative)** * **Procedural (skills)** * **Emotional conditioning** * **Priming effect** * **Conditioned reflex** ## General knowledge * General knowledge is culturally valued knowledge communicated by a range of non-specialist media and encompassing a wide subject range. * It is thought to be an important component of crystallized intelligence and strongly associated with general intelligence, and with openness to experience. * It is stored in semantic memory. * It is very important for making sense of the world and for learning new information. * Substantial amount of general knowledge related to knowledge of language (words and concepts). ## Semantic memory ### What is semantic memory? * General knowledge like ‘Paris is the capital of France’. * Knowledge about language and words - an average adult will know up to 50,000 words as well as lots of information about what each word means. * Other conceptual knowledge - a bicycle has wheels. All this information needs to be organised in some way in order to allow us to process information deeply and to search efficiently. Lots of our general knowledge is about individual elements - like dogs or cars. Using the idea of concepts and categories is a way that we can organise this information. ## What are concepts? * A category is a set of objects that belong together (birds, dogs, fruits). * A concept is a mental representation of a category of objects, typically derived from experience. * They are an efficient way of representing our knowledge of the world and help us to focus on the similarities among objects. * They allow us to make predictions about the world. * We also use concepts in communication to convey information about ourselves and the world - but there can be individual differences in beliefs about category membership. ## How do we define concepts? ### How do we categorise things? * How do we know that both rottweilers and poodles are dogs? * Initial explanations of concepts relied on the notion of ‘essential, common features’. * For example, dogs have fur, paws, tails, they bark etc. * Essentially the concept of ‘dog’ is an answer to the question ‘what is a dog?’ * Your answer is your concept of dog. * All dogs go into the category of ‘dogs’. This is problematic though as some concepts defy definition in this way. E.g. Wittgenstein’s problem of defining things like ‘games’ and ‘sports’. Some concepts have fuzzy boundaries, where they cross over with others (like pumpkins and tomatoes), so defining them by essential features is not really workable. Membership of a category is gradual rather than an all or nothing issue. So how do we do it…? ## Prototype model (Rosch & Mervis, 1975 - Hampton, 2010) * A prototype is an abstract idealised representation of the category member. * It is based on the average of members commonly experienced - e.g. robin, sparrow, starling. * Thus, it is like many members, but it is not exactly like any of them. * There are characteristic features, they are not necessary, but they are commonly present for many members of the category (e.g. flies for birds). * Items in any given category differ in their prototypicality - a robin vs. a penguin. Not all birds resemble the prototype though - emu, ostrich, owl, kiwi, penguin, etc. - they are birds, but atypical ones. Rosch described these as differences in typicality - robin high typicality, emu low typicality. Rosch (1975) gathered typicality ratings - a robin would be rated as very typical, an emu, not, a chair was rated as typical furniture, a telephone or mirror, less so. This supported the notion of gradual category membership. ### Prototype approach * Allows for fuzzy boundaries - if an object shares only a few features with the prototype then it likely shares features with another prototype - e.g. dolphins are fish (they are not, of course). * People offer ‘typical’ features rather than essential features - for example, ‘birds fly' even though some do not, but it is a feature shared by most category members, and is, thus, prototypical (Hampton, 2010). * Statements about prototypical objects are rapidly verified - ‘a robin is a bird' is verified quickly compared to 'a penguin is a bird' or 'an apple is a fruit' compared to 'a tomato is a fruit'. * This is the typicality effect. * Not all members have equal status (Meridan & Smith, 1984). ## Family resemblance * An object can be classified as belonging to a category when there is a match between it and the prototype. * We call this ‘family resemblance’, members sharing attributes/features. * Rosch & Mervis (1975) reported good evidence for the importance of this notion. * 5 most typical members of vehicle category - car, truck, bus, motorcycle, and train had 36 attributes in common. * 5 least typical members of vehicle category - horse, blimp, skates, wheelbarrow, and elevator had 2 attributes in common. * Also, there are high correlations between typicality and ‘family resemblance’ scores - family resemblance scores are the number of objects that share any given features. However, some categories show little evidence of family resemblance - goal-derived categories, like ‘things to take on holiday’ - passport, suntan lotion, holiday hat, book, tickets, phone charger etc. - these things have no family resemblance at all, and yet they are still grouped together (Barsalou, 1985). Barsalou also points out that people can categorise things into newly defined categories - even though this has no prototype to use for comparison. Hamilton (1981) points out that some categories do not have clear prototypes, concepts like justice, belief, etc. however… “Friendship” prototype * There is some evidence that we have prototypes for such ‘abstract’ concepts. * Fehr (2004) found that there is a prototype for “friendship intimacy” in same-sex friendships. Prototypical features of friendship intimacy: * Self-disclosure * Emotional support * Loyalty Less prototypical features: * Shared activities * Practical support ## Exemplar approach * According to the exemplar approach, instead of using an abstract ‘average’ description (e.g. prototype) we make category judgments by comparing new stimuli with instances already stored in memory. * The instances stored in memory are the “exemplars”. * New exemplars are compared with known exemplars. * The most similar item will influence classification the most. * More able to account for tricky categories like ‘games’, because actual exemplars, not prototypes are used, and atypical examples do not get ‘lost’ in an ‘average’. * Accounts for typicality - objects that are classified faster are like more of the stored exemplars. * Prototypes and exemplars may be combined and vary on their level of ‘abstraction’. ## Knowledge-based approach * In addition to knowledge of concepts’ features/attributes we also have an understanding of the relationship between the features/attributes e.g. sharp → dangerous. * Our knowledge of concepts is also influenced by these relationships, not just our knowledge of their attributes, and this has been shown to impact whether objects belong to certain categories (Ahn et al. 2000). * This is especially the case for casual relationships - i.e. members of a category have blurred vision, leading to headaches which leads to insomnia. * Participants considered those who did not have blurred vision as less likely category members than those without insomnia because of the causal relationship - if the cause is missing then category membership drops. ## Using concepts * The theoretical models and approaches considered so far treat concepts as abstract symbols and as relatively stable (e.g. a person uses the same representation of a concept on different occasions). * E.g. a vehicle consisting of two wheels, held in a frame one behind the other, propelled by pedals and steered with handlebars attached to the front wheel. ## The situated nature of concepts (Barsalou, 2003, 2008, 2009) * Representation of any given concept will vary from situation to situation depending on our current goals and important aspects of a situation. * We need to focus on the connection between conceptual, perceptual, and motor/action systems. * There is reasonable evidence to suggest that we access perceptual/situational and motor/action information when we process concepts, there is more to them than simply an abstract representation of the concept itself. ## Embodiment of concepts example * We now know that tongue, finger, and foot movements produce different patterns of activation in the motor cortex. * Hauk et al. (2004) presented participants with words such as “lick”, “pick”, “kick”. * These words activated parts of the motor cortex that correspond to the body part that would perform the movement. * So, activation of the concept causes activation here too. ## Hierarchy of concepts * Research by Rosch and colleagues has identified 3 levels in the hierarchy of concepts. * **Superordinate level** * Furniture, animals, tools * **Basic level** * Chairs, dogs, screwdrivers * **Subordinate level** * Desk chair, collie, phillips screwdriver ## Spreading activation model * Collins and Loftus (1975) suggested semantic memory is organised on the basis of semantic relatedness (semantic distance). * Whenever we see, hear or think of something the appropriate “node” in semantic memory is activated. * That activation spreads strongly to closely related concepts and weakly to distantly related concepts. * Explains the typicality effect - it takes less time to decide that a typical member belongs to a category (e.g. apple - fruit) than a less typical member (e.g. kumquat - fruit). ## Schemas * Schema is a mental structure we use to organise and simplify our knowledge of the world around us. * We have schemas about ourselves, other people, mechanical devices, food, and in fact almost everything. * They are larger knowledge units than concepts. ## Types of schemas: * **Social schemas** are about general social knowledge. * **Person schemas** are about individual people. * **Self-schemas** are about oneself - we also hold idealised or projected selves, or possible selves. * **Role schemas** are about proper behaviours in given situations. * **Event schemas** (or scripts) are about what happens in specific situations. ## Scripts * A narrower term than schema and refers to a well-structured sequence of events associated with a particular activity. * They help us to understand everyday events. * Example - lecture script- entering the room, finding a seat, sitting down, getting notepad, listening to lecturer, taking notes, checking time, leaving. * Scripts could interfere with memory by replacing expected events that were missing from stories. * Social structure encourages a strong degree of behavioural scripts to be utilised within everyday interactions with others. ## Schemas and memory * Schemas help us to understand situations - we know how to behave in certain situations because we have a schema for the situation. * They allow us to form expectations and make predictions about a situation. * They can influence our memory of an event as we are more likely to notice things and events that are congruent with a schema. * They help us prevent cognitive overload. * But they can cause various errors in memory and perception. ## Brewer and Treyens (1981) - Memory for objects in a room * Participants were 30 university students, who arrived individually to the laboratory and were asked to wait in an office. * After waiting for some time, participants were taken out of the office and asked to write down everything they could remember from the room. ### Results * Most participants recalled the schematic objects (desk, typewriter). * Some participants reported things that would be expected in an office but were not present (phone, books). * Many participants also recalled the skull. ## Stereotypes * Stereotypes - simplified generalisations about groups, are also a type of schema. * They can influence not only our social behaviour but our cognitive processing as well. * For example, there is plenty of empirical evidence that we often form impressions about people based on their accents. ## Studying stereotypes * Explicit measures (e.g. questionnaires) are not always reliable because they can be influenced by social desirability bias. * That is why implicit measures such as the Implicit Association Test (IAT) have been developed. ## Stereotypes - Gender Schema * Kreiner et al. (2008) asked participants to read sentences and recorded their eye movements while they were doing the task. * A. Yesterday the minister left London after reminding himself about the letter * B. Yesterday the minister left London after reminding herself about the letter * Participants fixate longer on a pronoun in sentence B * Similarly, when asked to review job applications identical in every way except for names, both male and female scientists preferred the male candidate. ## Stereotype stability * Research suggests that stereotypes are less stable and invariant over time than originally thought. * Garcia-Marques et al. (2006) asked participants to select 5 out of 43 traits that best describe various groups, then repeated the same task 2 weeks later. * Results reveal considerable variation between the traits selected in test 1 and test 2, especially for traits that are judged to be less typical. * Stereotype activation is flexible and dependent on context. ## Summary * Concepts are important constructs that aid our interaction with the world. * There are various explanations as to their nature and usage. * Many theories ignore how concepts are actually used. * Schemas and scripts are broader pockets of knowledge than concepts. * They can prevent cognitive overload, but can also lead to some errors. * Stereotypes are a specific type of schema.