Cognitive Development – The origins of abstract concepts

View Segments Segment :

  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
Successfully saved clip
Find all your clips in My Lists
Failed to save clip
  • Transcript
  • Transcript

    Auto-Scroll: ONOFF 
    • 00:03

      [Cognitive Development]

    • 00:09

      SUSAN HESPOS: Hi My name is Sue Hespos. [Susan Hepos, PhD,Associate Professor, Department of Psychology,Northwestern University] I'm an associate professorin the Psychology Department at Northwestern University.And today, I'm going to be talkingabout cognitive development.In particular, I'm going to talk about whatI think is one of the coolest things that we humans do.I'll be discussing the uniquely human ability to reason

    • 00:31

      SUSAN HESPOS [continued]: about truly abstract concepts.So think about the infinite sequence of natural numbers.You can't touch it.You've never seen it or tasted it.But, most school-aged children understand the concept of it.It's highly abstract.And they get it, even though they can't touch it.

    • 00:53

      SUSAN HESPOS [continued]: And my dog doesn't get it now, and he never will.So this capacity, I believe, comes from our abilityto use knowledge from separate domains,like number and space and objects and people,and combine these different domains of information

    • 01:13

      SUSAN HESPOS [continued]: to create new knowledge.So the ability to draw relational comparisonsbetween, say, objects, events or ideas.This analogical processing abilityis the key capacity supporting higher-order cognition.And I think it's what differentiates human cognitivecapacity from that of other primates.[Presentation Topics]So humans are unique in their ability

    • 01:35

      SUSAN HESPOS [continued]: to reason about truly abstract concepts.This analogical ability-- making comparisonsbetween ideas, events-- is the focusof what we're going to talk about today.In research on analogical processing,there are certain signatures to the abilitiesfound in children and adults.For example, relational learning in childrenis facilitated by seeing multiple exemplars

    • 01:56

      SUSAN HESPOS [continued]: and allowing the participants to make comparisonsbetween these things.So this is a way that we can facilitate the abilityto make a relational comparison.There's also ways that you can hinder the abilityto make a relational comparison.You can focus on an individual object instead the pairof objects as a unit together.

    • 02:17

      SUSAN HESPOS [continued]: And that can hinder learning.So we're conducting studies to see if these signatures thathave been of analogical reasoning, that have been foundin adults and children, are available in infants, as well.[Structure Mapping]

    • 02:38

      SUSAN HESPOS [continued]: So, structure mapping theory provides a mechanismfor explaining how we detect relational mappings.If we have two entities, depicted hereon the top left by these yellow hierarchies,we can draw comparisons about the relational structuresbetween these two things.So the idea here is with the top yellow node,you could draw a comparison and say, well,

    • 02:59

      SUSAN HESPOS [continued]: it's a lot like the top one on this one next to itand draw comparisons between the bottom leftand the bottom right nodes, as well.So this is highlighting the relational structuresthrough comparison.And this could be giving you candidates for inferences.So, for example, if you're a kid,and someone has labeled the one on the right as a horseand you're trying to figure out if the one on the left

    • 03:21

      SUSAN HESPOS [continued]: is also a horse, you might do a comparison between the shapeand size and recognize that they're closely alignedand draw the conclusion that they're both horses.So this is a really concrete thingto say the horses look identical or something like that.But it can also be a very abstract as well,such as comparing an atom to a solar system.

    • 03:42

      SUSAN HESPOS [continued]: They have object correspondences, like the sunto the nucleus and the planet to the electron.They also have common relational structures,like the planet revolves around the sundue to an attractive force between them.Likewise, the electron revolves around the nucleusdue to an attractive force between them.[Development Evidence for Relational Learningin Children]

    • 04:05

      SUSAN HESPOS [continued]: These are the concepts that we want to test.How has this been tested?On this next side, we're going to review two different tasksthat have been used to assess analogical learning abilities.[How has it been tested?]The first one is relational match to sample.So this is where you give the children a target.So it could be a picture of two pigs, a picture of two horses,

    • 04:26

      SUSAN HESPOS [continued]: anything like-- let's just talk about it as AA.So that's their target.You present them with AA.And then you give them a choice between two other ones,like XX or YZ.And you say, which one is more similar to the target.Which one is more similar to the first one you showed them, AA?And, if they're making a relational match,if they're given AA, they should use XX,because they're both the same.

    • 04:48

      SUSAN HESPOS [continued]: However, if they're given BC, theyshould go for YX, because they're both different.So this relational matches samplehas been used in a variety of speciesas well as a variety of adults and children.A slightly easier task that doesn't require languageis called the same-different task.And this is where you just familiarize by showing a baby,

    • 05:08

      SUSAN HESPOS [continued]: say, a pattern of two objects that are the same like AA, BB,CC.And then, after they've gotten habituated or a little bitbored with that, then you present themwith two test items, YZ or XX, and see which one they choose.And the same thing, you could familiarize themto a different pair and you'd expect the opposite pattern.These abilities, like I said, have been tested with children

    • 05:30

      SUSAN HESPOS [continued]: with a relational matches sample taskand with a variety of non-human primates,with some of them get the relational sample task,some of them do the same different task.So, it's interesting to note that childrencan do the relational matches simple task as youngas three years of age.

    • 05:51

      SUSAN HESPOS [continued]: Non-human primates-- and even bees--were successful in learning these relations as well.The difference was that it took the children very few trainingtrials to be successful.But it took other species, like bees and crowsand non-human primates, hundreds and sometimes thousandsof training trials, with feedback,before they could abstract this relation of same or different.

    • 06:15

      SUSAN HESPOS [continued]: So, the difference between humans and other animalsraises the question of whether or not this initial abilityis just a difference of degree.Do you need a few more trials?Or, is it a fundamentally different cognitive mechanismthat underlies this ability in humans?And there's no doubt that the influences of languageand culture facilitate detecting relational structures

    • 06:36

      SUSAN HESPOS [continued]: in humans.One way to address the question of if thisis a uniquely human aspect of our cognitionis to look at pre-verbal infants to seeif they look more like older childrenand learn the task with very few trialsor if they're more like non-human primates,and need extensive training.Let's first look at some of the signaturesof the analogical learning ability.

    • 06:56

      SUSAN HESPOS [continued]: So as I mentioned before, the ability to detect a relationis facilitated by having multiple comparisons.So if you were to present a kid or infantwith multiple exemplars, like AA, BB, then repeat AA again.And then say CC.And then give them, after all this training

    • 07:17

      SUSAN HESPOS [continued]: with multiple exemplars, you present XX and YZ.They're much more likely to make the relational matchand choose XX over YZ.Another thing that has been shown to hinder performancein terms of ruining their abilityto perceive the relation is if there's an object match.

    • 07:40

      SUSAN HESPOS [continued]: So what I mean by that is if you present AA,and then in the options that they have at test,you have XX and AZ.Younger kids tend to say, oh, there was an A in the first oneand there's an A in the second one.And so they choose the AZ match in that case instead of XX--so, instead of seeing the relational structure of same.

    • 08:01

      SUSAN HESPOS [continued]: [Do infants have an analogical process by which they canabstract relations?]Our question today is, do infantshave an analogical process by which theycan abstract relations?How early can we find this ability?So if they have this ability to abstract relations,does this process have the same signatures of analogical

    • 08:25

      SUSAN HESPOS [continued]: learning as in older children?So, are younger infants helped by multiple exemplars and arethey hindered by an object focus?[Signature of Relational Learning]And that's going to be the studies that we show you today.So, as we said before, signaturesof relational learning are when the relation is easierto detect because you've given multiple exemplars.

    • 08:48

      SUSAN HESPOS [continued]: And, it can be disrupted by object focus.So, if you had AA and presented AZ,they're more likely to make a comparison from Ato A than the relational match of XX.OK.So this is how we ask babies these questions.We tested seven and nine-month-old babies.And what we did is in this type of paradigm--

    • 09:10

      SUSAN HESPOS [continued]: it's called a habituation-dishabituationparadigm.[habituation dishabituation paradigm]Infants are seated on their parent's lapin front of a puppet stage.There's a video camera underneath the stagevideotaping their face and playing it for a researchassistant in another room.When the infant is attending to what's going on on the stage,the research assistant in the other roomis pressing a button on the computer.When the infant looks away from the stage,

    • 09:31

      SUSAN HESPOS [continued]: the researcher let's go of that button.If the kid looks away from the stagefor more than two seconds, that'staken as universal baby language for, I'm bored,get on with this.And the computer will beep, and the experimenterwill lower the screen and go onto the next trial.By showing babies different events on the stage,we record how long they looked at the eventsor how fast they get bored.

    • 09:51

      SUSAN HESPOS [continued]: And then after they've seen these habituation trials,we present them with test trials and see.Do they generalize their boredom and do short-looking timesbecause they think it's the same old thing over and over again?Or, do they recapture attention and go, wait a minute,this is something new?So what we did here is we presented them--half the infants saw trials with objects thatwere the same at the beginning.

    • 10:12

      SUSAN HESPOS [continued]: So as you can see here, there's some pink blocks.And that was shown on the first trial.And then there was two camels.That was shown on the next trial.Then red blocks, and then Elmo.And this is repeated.These same objects are repeated over and over again,from trial to trial, until the infant showed 50% declinein looking.After that, we present them with test trials.In the test trials, we alternate between showing them

    • 10:32

      SUSAN HESPOS [continued]: pairs of objects that are the same, justlike they saw previously.So they may say, oh yeah, been there, done that.And they'll have a short-looking time.And then, we alternate that with trialswhere the objects are different from each other.So on the top row, you can see the little red foam blocknext to a pig.Would they look longer at that after seeingseveral trials that were the same prior to that?

    • 10:54

      SUSAN HESPOS [continued]: So, if they looked longer at the different trialsduring test trials, we can concludethat infants learned the relationand detected when we changed from same to different.The other half of the infants sawpairs that were different at the beginning.And they saw a pink block next to a red block, the Elmo

    • 11:14

      SUSAN HESPOS [continued]: next to the camel.They were shown those on multiple trialsuntil their looking time declined.And then in test trials, we showed them the exact same testtrials as the other babies.And we'd expect an opposite pattern of looking.Let me show you the results.As you can see here, we've got trials along the x-axison the bottom.And along the y-axis, we have looking time in seconds.And during the habituation trials, just as predicted,

    • 11:36

      SUSAN HESPOS [continued]: the babies get bored with whatever they'represented over and over again.So if they saw the relation was samethat they were habituated to at the beginning,they looked for a long time, about 40 seconds,on the first trial.And then they get down to about 10 secondsby the end of the habituation sequence.Then we bring them over into test trials.And here, I have flaps across babiesthat saw it same and different.

    • 11:57

      SUSAN HESPOS [continued]: So in the test trials, it was what was the novel relations.So if they saw same at the beginning,did they look longer at different?If they saw different at the beginning,did they look longer at same?And what this graph shows at the testtrials is that infants look significantly longerat the novel relation.So they were able to detect this abstract relation.So now, let's dig down a little bit deeper.

    • 12:19

      SUSAN HESPOS [continued]: These are the same pictures that I showed youbefore showing what they saw during habituation the sameand habituation the different.Now, what's really neat about this first pair of test trialsthat the kids saw was that if you look at all the pictures,they're brand new objects.So this is like the gold standardfor testing whether or not kids reallyabstracted the notion of same or different

    • 12:39

      SUSAN HESPOS [continued]: because we showed them the relation of samewith pink blocks and red blocks and Elmo and camels.But if they really abstracted the relation,they should be able to apply that relation of sameonto a brand new set of objects that they haven't seen before.So like the one on the left with the green puff balls on itwas same, and they had a low looking timefor that because they were bored with it.

    • 12:60

      SUSAN HESPOS [continued]: But they looked significantly longerat the red foam block next to the pig.OK.So if we just look at that trial alone,infants looked significantly longer at the novel relation.So that's pretty neat because they're definitelyabstracting this relation.Now, another thing we did is we wanted to know whether or notthis object focus or object experience wouldruin the ability to perceive the relation.

    • 13:20

      SUSAN HESPOS [continued]: So prior to the experiment, now depictedalong the top of the screen, we showed babiesindividual objects.So they're not in a pair relation.They're just a single pink block-- a single dragon,a single Elmo.And we give five seconds of experience just looking at it,tapping on it.And then we give them another object, looked at it,tapped at it.And so we showed them a variety of objects.

    • 13:43

      SUSAN HESPOS [continued]: Now, as you can see on the top, there'sthis little blue character with a big nose.And he's presented in the line along the top.They didn't see that blue guy in the habituation trials at all.So they only saw the individual in the waitingroom prior to the experiment.They didn't see it during the habituation training.And then the next time they see it is during test trials.In this case, infants were unable for these objects

    • 14:05

      SUSAN HESPOS [continued]: to learn the relation.We think what happens is that theysee the blue guy in the waiting room.They're like, there's a blue guy, he's really cute,he's fuzzy, he's neat.And then they don't see him again until test trials.And they're like, I love that blue guy.And they never recognize whether or notit's in a relation of same or different.And so for those particular tests trials,they don't learn the relation because the object

    • 14:29

      SUSAN HESPOS [continued]: focus has interfered with their ability to learn in this case.And then finally, in the last case,we wanted to see how flexible are these guys.So if you notice in this one, during the test trials,there are pink blocks in this particular last caseon the bottom.They saw the pink blocks in the waitingroom prior to the experiments.

    • 14:50

      SUSAN HESPOS [continued]: So they've got that individual experience with it.But then-- so that should ruin their abilityto learn-- but then in habituation trials,we also show them the pink block next to another pink block.And that gets them over this object focus.And they go, oh, it's next to that other one.They draw comparisons between themand it remediates them so that in test trials when they're

    • 15:10

      SUSAN HESPOS [continued]: presented with this again, they can detect the new relation.So again, taken together, this detailed analysisof what's going on tells us that the signaturesof relational learning are available that we'veseen in adults and in children, they alsoseem to be evident in seven and nine-month-old infants,

    • 15:31

      SUSAN HESPOS [continued]: as well.[Conclusion][Summary]Alignment of multiple exemplars facilitatesrelational abstraction.What's also interesting about these resultsis that we had equivalent performancefor babies that were habituated to different or habituatedto same.Many of us have an intuition like, wow,

    • 15:51

      SUSAN HESPOS [continued]: detecting that two things are the same ismuch easier than detecting two things are different.When I walk down the street and I see two identical twinswalking next to each other, I go, wow,those guys look the same.But I don't spend the rest of my lifewalking down the street going, boy,that person looks different than that person, than that person.And so, we have this intuition like,

    • 16:12

      SUSAN HESPOS [continued]: maybe it's easier to learn same prior to different.But at seven and nine months of age,there's no difference in the learningthat it takes to abstract either of these relations.So that wasn't supported by the data.But it's still interesting.[Summary]We've also found from these data thatobject focus disrupts relational learning.So that's when they saw it the waiting room.They didn't see it during the habituation trials.

    • 16:33

      SUSAN HESPOS [continued]: And the next time they saw that blue guywas during the test trials.They were unable to perceive the relationthat that blue guy was part of.And so that's a way of interfering with learning.And that has been shown in children as well as in adults.So taken together, seven-month-oldsshow two signatures of relational learning

    • 16:54

      SUSAN HESPOS [continued]: found in older children.So we uncovered the earliest evidence.So taken together, what these findings seem to showis that we uncovered the earliestevidence of our cognitive capacities in infancyand described what changes over development.Through this process, we gain informationthat's critical to understanding cognition in general.So the patterns revealed in our researchwill give parents and educators requisite knowledge

    • 17:15

      SUSAN HESPOS [continued]: to support learning.In addition, our findings will enable practitionersto better identify patterns that deviatefrom typical targeted intervention for childrenwith cognitive impairments.[Further Reading]

Cognitive Development – The origins of abstract concepts

View Segments Segment :


Professor Susan Hespos describes her research into the abstract reasoning skills and cognitive development of infants. Her team performed experiments to determine whether babies tell the difference between matching and nonmatching pairs of toys. They concluded that babies do have the capacity to do this, but that it could be disrupted by object focus.

SAGE Video Tutorials
Cognitive Development – The origins of abstract concepts

Professor Susan Hespos describes her research into the abstract reasoning skills and cognitive development of infants. Her team performed experiments to determine whether babies tell the difference between matching and nonmatching pairs of toys. They concluded that babies do have the capacity to do this, but that it could be disrupted by object focus.

Copy and paste the following HTML into your website

Back to Top