LANGUAGE, THINKING AND MEMORY

1
LANGUAGE, THINKING AND MEMORY

• INTRODUCTION

2
Main Topics

• Language (Alan Garnham)
• Thinking (Alan Garnham)
• Memory (Sam Hutton)

3
Language - Lecture Topics

• Word Meaning
• Spoken Word Recognition
• Written Word Recognition and Reading
• Sentence Processing
• Text Processing
• Language Production and Speech Errors

4
LANGUAGE, THINKING AND MEMORY

• WORD MEANING

5
CONCEPTS, CATEGORIES, AND WORD MEANINGS

• Concepts define categories (of objects, actions,
  or whatever)
• Concept is intension (with an s) or criteria
  for membership of category
• Extension is set of category members
• (Some) concepts are lexicalised
• They have a single word associated with them

6
WORD MEANINGS

• Important in talking about (categorised) things
  in the world
• Word meanings can be combined (in sentences etc.)
  to convey complex pieces of information

7
MAJOR TYPES OF "PHYSICAL OBJECT" CONCEPT

• 1. NATURAL KIND TERMS
• Things that exist in the natural world types of
  animal, plant etc.
• The "true" reason for things of that type being
  alike (e.g. what all dogs really have in common)
  may have to be (at least partly) determined by
  scientific enquiry.
• e.g. dog, rose, water

8
MAJOR TYPES OF "PHYSICAL OBJECT" CONCEPT

• 2. ARTEFACTS
• Things people have made to serve a particular
  function, and which are therefore defined in
  terms of their ability to fulfil that function.
• e.g. chair, car, lubricant

9
OTHER TYPES OF CONCEPT

• Relations
• Verbs
• kick (A kicked B)
• put (A put B on C)
• Prepositions
• above, under, on, in
• Properties
• Adjectives (red, good)
• Adverbs (slowly, suddenly)

10
PROPERTIES OF CATEGORIES

• Some categories have rules governing their
  membership (e.g. even number)
• Many categories have more and less typical
  members (bird robin vs ostrich)
• Typicality gradient or graded structure
• Many categories do not have clear boundaries (are
  book-ends furniture?)

11
CATEGORY HIERARCHIES

• Object categories often form hierarchies
• Edibles -gt Fruit -gt Apple -gt Cox
• Distance apart in hierarchy can affect judgement
  time
• A robin is a bird easier than
• A robin is an animal
• (e.g. Collins and Quillian, 1969)
• The feature model of Smith, Shoben Rips
  accounted for much the same data
• More specific concepts (dog) have more features
  (vs. animal)

12
CATEGORY HIERARCHIES

• Rosch identified a basic level in such hierachies
  (e.g. apple) together with superordinate (fruit)
  and subordinate (Golden Delicious) levels
• Items at the basic level
• Have many attributes (apple vs fruit)
• Have attributes distinct from other basic level
  categories (apple vs banana)
• Are named quickly
• Are learned early by children

13
CONTEXT DEPENDENCY

• Crucial attribute of a concept vary from context
  to context
• E.g. value vs hardness of a diamond (selling vs
  cutting)
• Concepts may be created on the fly
• Barsalou ad hoc concepts such as items to be
  rescued from a house on fire

14
WHAT IS A CONCEPT?

• Necessary and sufficient conditions for class
  membership (the classical view)
• Things that are more or less like a prototypical
  member (the prototype view)
• A set of instances (the exemplar view)
• A theoretically-based method of linking a set of
  things together (the theory view)

15
THE CLASSICAL VIEW

• Bachelor unmarried male adult
• Does not account for typicality effects
• Category membership usually thought of as all or
  none
• Hard to give clear cut definitions (the defining
  features should provide these)
• Wittgensteins notion of family resemblance
• Does not, of itself provide constraints on
  combinations of features

16
THE PROTOTYPE VIEW

• Characteristic attributes or best example
• Focal examples of colours
• Robin for bird
• Explains typicality effects and lack of clear
  boundaries (how like the prototype does the
  category member have to be)

17
THE PROTOTYPE VIEW - PROBLEMS

• Some attributes predict category membership
  better than others
• Does not account for what makes a good concept
  and what does not
• Concepts are not just the sum of constituent
  attributes

18
THE EXEMPLAR VIEW

• Can explain typicality effects - typical examples
  of a category are like lots of previous examples
• Examples directly reflect variability within a
  category
• Like the prototype theory depends primarily on
  similarity

19
THE THEORY THEORY

• Theory or explanatory framework explains why
  things are grouped together
• Natural kind terms
• Clean and unclean animals (for eating or not)
• Does not depend purely on similarity, but on one
  type of similarity rather than another
• Children use knowledge rather than feature
  similarity to judge category membership (Carey
  and others)
• Markman and Gelman (1987) owl grouped with
  flamingo rather than bat

20
CONCEPTUAL COMBINATION

• A fake gun is not a gun
• An expert repair is not a repair done to an
  expert (rather one done by an expert, cf. Engine
  repair)
• A prototypical pet fish is not a prototypical pet
  or a prototypical fish
• Small spoons are more typical than large spoons,
  metal spoon are more typical than wooden spoons,
  but large wooden spoons are very typical (Medin
  and Shoben, 1988)

21
CONCEPTUAL COMBINATION

• The knowledge based view of concepts can help to
  explain conceptual combination.
• An expert is someone likely to carry out a repair
  an engine is something likely to be repaired
• Hence the difference between expert repair and
  engine repair
• Murphy (1990) predicted and found that NOUN-NOUN
  combinations would typically be more difficult to
  understand that ADJ-NOUN combinations.