Chapter 9 Lecture Flashcards
Conceptual knowledge
Knowledge that allows us to identify objects and events and to make guesses about the objects nature
Concept
A mental representation, (a depiction of something in our mind) that is used to carry out many cognitive tasks/procedures, (cognitive functions)
Categorization
When things are split into categories- a category includes all exemplars of a concept so all examples of a mental representation used in a variaty of cognitive process?
Why are categories important?
Allow us to know the general and potential specific features of an item making it so thta we can know that an item is part of a group based on these features even if we have never encountered the item before
What is the definitial method of categorization
Goes based on the definition of a category, does not work well as many members of the same everyday chtegory can have different defining features - under the definition approach everything should perfectly match the definition which can be hard/forced to use unspecific definitions that could also appply to toher things
What is the prototype approach to categorization?
We make the standard that we use to determine if somethign belongs in the category or not based on degree of simmilairty to the standard. Under the prototype approach the standard is the average of all members of this category we have previously encountered.
ex if forming a prototype of three boats our prototype boat would contain features from all 3 and thus would be a combination of all the ones we have previously encountered but in itself not one that we have previously encountered.
Describe the prototype coglab
Study phase
Presented with an image of a pattern of dots and asked to select weather it more closely resembles pattern A or pattern B, (each image included the same number of dots having been moved the same number of spaces from their original position in their original image- however exact pattern of dots and which directions were moved in varied from image to image) Participants have never seen pattern A or B so in this phase they just guess and get a sense what pattern b must be like based on the images that they were told were distortions of pattern b and the ones told were distortions of pattern A. Researchers theorized that participants would take the average of all of the distortions that they were told were distortions of pattern to form an imagined image of pattern A and vice versa with pattern B.
Test phase
Showed participants one new distortion of pattern A and the original pattern which all of the previous distortions of pattern A the participants had encountered had came from. believed that if participants had come to represent pattern
a as a version of the average of all previous distortions then this averaged image would be most associated with pattern A in their mind and would likley be close to the actual prototype of pattern A, therefore they would likely be able to recognize the prototype of pattern A much more quickly then the new distortion. Researchers did find that participants had much shorter reaction times when identifying the original patterns so proved that particiaptns had formed prototypes in their minds.
What is the prototype coglab based on
Posner and Keele 1968
participants were shown 12 different images that could be distortions of a few different potential original images same as cog lab when first asked which pattern the image was a distortion of they had to guess, (as they had never seen the pattern before). Participants were then told which pattern the image was actually a distortion of, this allowed for participants to mash together all of the images they had seen before and were told were distortions of a particular pattern to get a sense of that pattern, *take the average of the distortions to get a sense of the original pattern).
Test phase
Participants were shown a distortion they had already encountered in the study phase and asked to classify which pattern it was a distortion of, (this should have been easy as they might remember exactly from the study phase which pattern it had been a distortion of and therefore they should have quick reaction times for classifying it), a new distortion, (one that they had not encountered in tthe study phase should be slightly harder to classify because it is not fammiliar and therefore should take slightly longer to classify), a far-disrotion (where the dots were moved more spaces then they had been in all previous disrotiots should also be hard to classify) and the prototype which reaaserchers therorized that if participants had formed a mental image of the average of all of the previous disrotiotions that they associated the most with a certain category they should be able to recognize it more quickly.
findings
founda that participants were able to classify the original image as belonging to a specific pattern group just as quickly as they were able to classify the distortion that they had previously seen- this very strongly suggests that they had formed a prototype using the average of all of the distorted images.
What does high prototypically mean?
An example of a category very closely resembles the average of all known members of that category ex. if the majority of birds that you know of are small birds that can fly then a robin will very closely resemble your bird prototypes flight abilites and size, (that you dervied by taking the average of these features from all the birds you know)
What does low prototypicality mean?
An example of a category does not very closely resemble the average of all known members of that category ex if the majority of birds that you know of are small birds that can fly then a penguine will not very closely resemble your bird prototypes flight abilities and size (that you derived by taking the average of these features from all the birds you know)
Are prototypes subjective?
Yes prototypes are subjective becasue they are based on the indivkduals prior knowledge
What was Rosch’s prototype experiment?
Gave participants categories ex. birds and then a list of 50 birds, (50 members of the category she gave) and told participants to rate on a scale of 1-7 how good examples of the category these members were. 1 meant that tney had high prototypciality, (that participants considered them to be good examples of the category and therefore to be very simmilar to other examples reflecting their simmilairty to their prototypes for that category, (the average of all members of that categories features)) and 7 for members that they thought had low prototypicality, (that were not very simmilar to others in the category therefore not simmilar to the prototype of the category and not good examples of the category)
found that for birds sparrows rated 1’s bats rated in high 6’s
found for furniture sofa rated in 1’s range and telephone rated in 6’s range
Describe Rosch and Mervis’s family resembalance experiment
Gave participants a list of 4 items from the same category- in this case furniture
ex. chair, sofa, telephone, and mirror
asked partiicpants to write down features they would typically see in each example
ex. chair
four legs, can be sat on, …. etc.
When two items had many of the same features they were said to have a strong family resembalance. The more of the same features an item in the category has with other items in the category the more likely those specific expressions of features are to be represented in the expression that comes from taking the average of all expressions of that feature for all items in the cateogory, therefore having lots of shared features with other items in the category results in higher prototypicality, (the item is more like the prototype) conversly having few shared features with other items in the cateogry makes an item have a low family resemblance to those items and means that it has low prototypicality- it is not simmilar to the prototype for this category
Describe Edward Smith’s 1974 sentence verification experiment
Presented participants with a sentence and measured how long it took them to respond yes or no to if the sentnece was true or not.
ex is this sentnece true or not: “a mirror is a fruit” - answere no, “an apple is a fruit” answere yes, “a pomegranate is a fruit” answere yes. Found that people responded more quickly when saying that an apple was a fruit then when they were saying that a pomegranate was a fruit. This is likely becasue an apple has many of the feature expressions that are expressed when determiining the average for each feature for the prototype of a fruit- apples have many shared features with other fruits so they are seen as good examples of fruits and therefore it requires little consideration if they are fruits or not because they match the prototype so well. Pomegranates or something like tomatoes do not match the prototype as welll so participants need more time to consider weather or not they are a fruit- is not instant match to protoype. If lisiting out features of all fruits, (objective) and comparing them apples have a greater family resemblance with most other items then pomegranates and tomatoes, (might not be fully objective because it would depend on what furits were listed unless all fruits were listed)
What position are prototypes given in category lists?
Mervis - found that items with high prototypicality are usually listed first when listing all members of a category, this is likely because they are more strongly associated with t
Describe Rosch’s priming experiment
One group ranked how good of an example of its corresponding color each shade was, (ex ranked a rich green as a good example of green and a pastel green as a poor example of green)
Then another group indpenedent, (seperate) from the group who did the ranking was given a task where they would hear a primer ex. the word green and then could be presented with two different colors and asked if the two different colors were the same as each other, ex if got poor green-poor green or rich green - rich green would say yes and if got yellow- blue would say no and measured the reaction time of their response. The idea was that when they heard the primer they would imagine a prototype, (what they based on all members of the category they had previously encountered consider to best represent all/the average of all) for that color. If the color they saw wasa the one that the other group had seperately rated based on how good of an example it was of that color or not was a good example it was more likely to be deemed more representative of the most posibilities of that color so it was likely simmilar to the prototype. If participants already hadd the image of the prototype and were rpresented a color that very closely matched it they already would have some information regarding the information they ad been given in their brain before hand making it easier to process- therefore colors that better matched the prototype would have quicker reaction times. Did find that reaction times for ones that more closely resembled prototypes were better.
What is an exemplar approach to determining if something fits into a category or not
Involves comparing item to each individual previously encountered item rather then the average of all previously encountered items.
Describe Posner and Keele’s 1968 bit distortion experiment
Training phase
Split participants into 2 groups
Group 1: Shown multiple distortions of 4 different prototypes where the dots are moved 1 position from where they were on the original image
Group 2: Shown multiple distortions of 4 different prototypes where the dots were moved 5 positions from where they were on the original image
Theorized:
Since all or most dots are moved the same number of spaces within a group participants should get a general sense of the dots position in relation to each other this should result in them both forming a prototype (their pattern). The prototypes should be very simmilar, the 5 space group should have essentially a stretched out version of the 1 space groups prototype, (as the average dot position in relation to other dots should be the same- especially if shown same positon changes, should just be on different scales). Overall if prototypes only were used to classify the patterns then the reserachers theorized the 5 and 1 space dot should have equal reaction times when categorizing new distotion images where the dots had been moved 7.7 spaces.
Findings: The group who had, had the 5 space distortion was able to classify the 7.7 space distorition image more quickly. 5 is closer to 7 so proves that simmilarity in scale not just general pattern, (prototype) also played in role in making classifications.
What is heirarchial organization
When a more general group can be broken down into smaller groups with more detailed criteria, creating different levels of broadness of category
ex. the larger group of chairs can be broken down into the smaller groups of dining chairs, living room chairs, etc.
the categoy of chairs would be the broadest level in theis ex and the smaller groups of dining chairs and living room chairs would be more specific levels in theis case
What are the different levels in hierarchal organization?
Superordinate or global- broadest level, ex furniture
basic level all the more basic subcateogries that can be made from furniture ex. chair, table bed
Subordinate or specific - most specific level
furniture
chair table. bed
dining, kitchen kitchen, living room single double
superoridnate/global, (superheros protect the greatest ammount of people possible if they protect the whole earth–> (globe) the superordinate or global category simmilarily refers to the broadest possible category that can be made in an example)
Basic - basic some people think is mid- middle level
subordinate or specific
can have subcategories of a category referred to as children
Which do people use to make classifications?
Exemplars are better for greatly varied groups and avoiding having atypical members get lost on average, (ex if have sparrow, robin, hummingbird, lark, raven, crow, geese and penguine and are creating a prototype based on average flight abilities you have some birds geese that can fly very long distances, lots of birds that can fly considerable distancesy the and penguines which can not fly at all if you were to try to determine the average distance that the birds on this list can fly then the penguines not being able to fly would avoid it having being skewed as very far by the geese but would still hav average distance able to fly would be decently far becasue you have very far and many decently far and one not, (not/very far kind of cancle each others effects) this would mean that when categorizing birds you may imagine that birds can fly and therefore leave out flightless birds) if you use an exemplar on the other hand even if you are given a bird dissimarl from most other birds you have encountered bc it is flightless if you have ever encountered a flightless bird before you can still know it is a bird by comparing it to that specific example.
We tend to use prototypes initially and then we strengthen our examplar knowledge later onlikely because we encounter members that don’t match much of protytpe that we still recognize as members of the group
Examplars may be better for smaller groups ex people who won a specific award could probably memorize all examples and then compare latest ones to them individually in your head however if it is a larger group like birds it is better to use a prototype because to many exist to go through them all individually in your head.
Describe Rosch’s privilieged level of categorization experiment
Gave pariticpants 3 trials
trail 1 asked participants to name all the features of furniture, (superordinate/global level in this ex) they could think of in 2 minutes
trial 2 asked participants to name all the features of tables, (basic level of this ex) that they could think of in 2 minutes
trail 3 asked participants to name all the features of kitchen tables (specific level in this ex) that they coudl think of in 2 minutes
the average number of the number of features named by participants for the superordinate/global level, (furniture) was 3, for the basic level was 9, and for the subordinate/specific level was 10.3
Fin that going from superordinate to basic level you gain lots of information, ( average was 3 features for the superordinate level compared with 9 for the basic level) and basic to subordinate only slightly increased information, (average number of features that could be named for the subordinate level was 10.3)
Which level is “special”?
The basic level is “special”
Rosch experiment
show participants pictures ex an image of blue jeans and ask them to as quickly as possible name the object in the picture the participants could have named at the superordinate level, (clothing), the basic level, (pants) or the subordinate level, (jeans)- found that participants were more likely to use the basic level name then the name from other levels.
Children usually first learn about concepts by learning their basic level name
Different cultures tend to divide living things into the same basic level categories, (ex have bird exist as a category across cultures/have an option of dividing animals by species- wonder if this would occur even if dont’ have broader or narrow categoreis)
Adults more likely to use basic level terms when talking then subordinate, or superordinate
Describe Quillian and Collins 1969 model
Based on a computer
start at the top node, (label) with the most general concept, (all nodes below this node are a subset, (children) of this node) then go then write the names of the second most general concept first children of top parent node, ( where you can divide all members of the first concept into the least categories possibel) by drawing a link between it and the general concept, (node) keep writing subsets and links until youu get to the most speciifc node that you are intereested in. Each descriptor is placed on the highest node in “family tree” that they can be while still being applicable to the majority of the children, (so descriptors are placed at the node at the most general level they can be) - the reason for this is to save ffort of writing descriptor for aeach individual child it is applicable to if it is applicable for the majoority of children, can instead then write it by the parent node, (which tells you that it applies to all children of the parent unless specified otherwise on their level). The number of links between one node and another node show how related they are to each other, how many nodes you have to go from one node until you get a link that connects to it and another node that if you go down enough’s children, grandchildren etc. will get you to the other node you are looking for. Ex shark and canary have to go up link from canary to bird- since neither shark nor one of its parent nodes is not under the node of bird have to go up to birds parent which is animal then have to go down to the child fish and then the grandchild shark, you have four links between them shows that they are less associated then say canary and another bird which just have two links.
