Lecture 18 - Types of Knowledge/Memory Flashcards
semantic memory
your knowledge base: the things/facts that you know
version of LTM
(i.e. memory for
categorical, factual information)
episodic memory
the events that you remember
things that happened to you or things that made an impression
procedural memory
skills
things you can do/know how to do
declarative memory
things you can state or verbalize
implicit memory
the things that you’re not aware that you know
stuff that you can do but you’re not consciously aware that you can do
explicit memory
stuff that you’re conscious of (when you recall something
Semantic memory is thought to
have three primary characteristics:
1) It is organized by content
2) It allows us to make inferences
3) Semantic info generalizes beyond a single episode
Semantic memory is organized by content
Similar concepts are more likely to cue one-another through association. [lexical decision task]
e.g. “doctor and nurse” are content that share alot of context so they’re more tightly associated than “doctor and table”
Semantic memory allows us to make inferences
There appears to be a logical, hierarchical order to the
categories (to LTM).
Semantic info generalizes
Information is not tied to a single instance but rather is more reflective of all the things that have happened
One way of testing semantic information is
using a lexical decision task (Meyer &
Schvaneveldt, 1971).
trying to decide: is something a word or not?
• A string of letters is presented (e.g. PRINTER or
PROGER) and you judge it is a word or nonword.
- The letter string is PRECEDED by a related or unrelated word.
- Related primes (e.g. nurse –> DOCTOR) were faster than unrelated primes (e.g. butter –> DOCTOR).
• What does the speed of retrieval say about the
internal structure of memory?
priming
sets up activation in a network
feedforward activation of some concept (nurse) that goes to activate other related concepts
Semantic memory is thought to be structured and allows
inferences.
- Is a canary a bird?
- Is a canary made of molecules?
• The first question requires retrieval from long-term memory; the second question is likely novel and requires an
inference or other knowledge - semantic relatedness (and takes a bit longer):
canary -> bird -> animal -> physical object -> made of
molecules
• The speed of retrieval and nature of the inferences suggests a possible structure to long-term memory.
The speed of retrieval (taking longer for novel objects) and nature of the inferences suggests
a possible structure to long-term memory.
what kind of structure does memory follow?
hierarchical
Collins & Quillian (1969) proposed a hierarchical model of long-term memory.
very much like a computer model: digital conception of what knowledge would ideally look like
an attempt to use logical operators
• Nodes in the model are concepts or ideas (e.g. bird, animal).
• Links are labeled (“is-a” or “has-part”) and directed. Subordinate categories
point to superordinate categories.
• Activation tags help verify inferences in the network.
− Canary is yellow (faster)
− Canary has skin (slower)
Nodes in hierarchal model
not just words but concepts
are concepts or ideas (e.g. bird, animal).
each node has links to other elements
Links in hierarchal model
are labeled (“is-a” or “has-part”) and directed. Subordinate categories point to superordinate categories.
links point in a direction, natural flow
sometimes they point to a property or the relationship between different nodes
“is-a” link
the canary “is-a” bird, a bird “is-a” animal
the way things are organized in the structure
“has-part” link
a bird has wings, has feathers, other qualities that that node (concept has)