Introduction and History:
- Disciplines in cognitive science
- Empiricism, nativism, behaviorism, functionalism
- Marr’s three levels
How is perceptual and cognitive processing, the
remembering of information, and so on, actually done with
neural hardware in the brain?
Often this is the focus of cognitive neuroscience
What processing steps are made to make a decision, or
produce behavior, or so on?
Often this is the focus of cognitive psychology
Why does the cognitive capability behave like it does?
What is its goal or purpose?
Often this is the focus of artificial intelligence or
Concepts and categories:
Concepts: mental entities
Categories: collections of stimuli in the real world
- Definitional, prototype and exemplar theories
o Definitional: (set-theoretic) approach assumes stimuli are grouped
using a set of necessary and sufficient properties. Does not work
for real-world domains.
List of conditions that need to be met for a stimulus to
belong to a category.
Example: triangles are closed shapes with three straight
o Prototype: Assumes people categorize stimuli by similarity to a
prototype, which is an “ideal” instance of the category.
Example: A bird is more likely to be a robin (more typical)
than an emu (less typical outliers).
Every instance (exemplar) of a category is remembered
New stimuli are categorized by the average similarity they
have to all category exemplars
- Schemas, scripts o Schemas: Example: kitchen schema, items that belong in a
structure, stereo typical, such as stove, fridge, sink etc. (Not like a
toilet, bed, lounge…)
Having ‘slots’ filled with ‘variables’
o Scripts: Schemata for events, rather than structures. Capturing the
stereotypical pattern, but allowing for constrained. Example:
Going to a restaurant and ordering a meal. Some flexibility and
- Ad-hoc and goal-derived categories
o Ad-hoc categories: Emphasize concepts that are not part of long
term knowledge structures but can be created ‘on the fly’ in
response to specific goals and circumstances.
Example: “things to take from a burning house”
Can become more permanent and well-defined through
o Goal-derived categories: Categories that are well established are
sometimes called “goal-derived”
Example: Apple as ‘snack food’
- The basic level
o Basic level is preferred level such as Chair verses furniture or
Perception, action, cognition
- Top-down apperception and bottom-up perception
o Top-down apperception: top-down, cognitively driven sources of
Memory, knowledge, concepts, ..
o Bottom-up perception: bottom-up, sensory driven courses of
information from external stimuli
Visual stimulation, auditory stimulation, ..
- Context effects of similarity
o This shows that when comparing two items whether stating the
differences or similarities, it makes them more similar in recall.
Ruling out any uninteresting possibility compared stimuli because
of a shared feature “things I compared”
- Change detection and change blindness
o People fail to detect large changes to visual arrays and scenes
if they are briefly occluded
o Change detection: The task
Example: Transparent video with women walking through
with an umbrella. Non-transparent of a gorilla walking
o Change blindness: the inability to perceive change
Example: building picture with more of the same picture
but missing something. People fail to notice.
- Perceptual illusions (visual, McGurk effect) o McGurk Effect: McGurk and MacDonald (1976) studied how visual
cues affect auditory perception
Presented an auditory [ba] sound being paired with the lip
movements for a [ga] sound
People perceive da [da] or [tha] sound
- Categorical perception
- Template matching, feature detection theory, and their applications
o Template matching: having infinite number of templates to
recognize a standard.
o Feature Detection Theory: Such as find T in a groupe of Z vs find T
in a group of Y.
- Perception as inference
- Embodied cognition
o Embodied cognition emphasizes intelligent agents being situated
in, and acting within an environment.
Embodiment makes it possible to think about non-
perceptual stimuli, by replying on spatial metaphors.
- Deductive and inductive decision making
o Deductive: Possible to deduce the correct answer
Example: Watson Task, or boat with animals
o Inductive: Human decision-making from reasoning from a specific
observation to a more general conclusion
Example: Which McValue meal to choose, what courses to
enroll in, etc.
- “Rational” and heuristic decision making
o ‘Rational’: describe human decision making in terms of
maximizing benefit or utility
o ‘Heuritic’: describe human decision making in terms of “rules of
thumb” or “approximate solutions”
- Heuristics and biases approach
o Representativeness, availability, anchoring and other heuristics
(law of small numbers, ignoring base rates, …)
Representative: people have the tendency to judge
probabilities or likelihoods accround to how much one
thing resembles another.
Example: Feminist bank teller
Availability: People assess the frequency of a class or the
probability of an event by the ease with which instances of
occurrences can be brought to mind.
Example: Much easier to image a death by shark,
much easier to image a word beginning with k, etc.
Anchoring: Decisions are disproportionately influences by
the first available pieces of information Example: solve math problem, because the largers
numbers were first, people believed that it was a
larger income vs placing smaller numbers in
Example: Estimation, when given a certain
percentage people would make a guess percentage
related to the given.
Law of Small numbers:
o Prospect Theory
Value function for gains is concave
Value function for losses is convex and relatively more
This implies that being risk averse for gains but risk
seeking for losses
Losses “loom large” and are weighted more heavily than
- Fast and frugal heuristics approach: Limited search and simple decision-
making because the world is competitive and resources are valuable, so
you need to be FAST! The worl is changeable , so you need the robustness
that comes from simplicity.
o Recognition heuristic, theory and applications
Recognition heuristic: When one of the two objects are
recognized, the recognized object has the higher value
Unless, the recognized object is negative then it will
have the lower value.
Example: Comparing cites!
“LESS IS MORE” Effect
if recognition heuristic is more accurate than
knowledge, partial familiarity will lead to the
- Correlated environments: where the fist pieve of information predicts
o Example: Real-estate agents believe the first pieces of information
(approaching house) determine wheater a viewer will decide they
want to buy the house
- Environments of diminishing returns: where the first pieves of
information are more important.
o Example: The starting give in backetball are much more important
than the bench in determining the outcome.
- Klein’s recognition-primed decision-making approach
o Do not need to compare options, can j