Barbara Landau: Does Language Change Thought?
Resumen
TLDRSusan Goldin-Meadow introduces Barbara Landau, a distinguished researcher in language and cognition. Landau's work explores how language influences thought, particularly in children, including those with disabilities. She discusses two main hypotheses: one suggesting that language shapes non-linguistic thought, and another proposing that having language transforms thought processes. Landau presents her 'recoating hypothesis,' which posits that language provides a formalism that enhances cognitive abilities without altering underlying non-linguistic representations. She emphasizes the momentary effects of language on cognition and highlights the importance of syntactic structures in shaping thought.
Para llevar
- 👩🏫 Susan Goldin-Meadow introduces Barbara Landau.
- 📚 Landau's research focuses on language and cognition.
- 🧠 Language influences thought, especially in children.
- 🔍 Two main hypotheses: language shapes thought vs. transforms thought.
- 🔄 The recoating hypothesis suggests language enhances cognition.
- 📝 Syntax plays a crucial role in shaping thought.
- 👶 Blind children learn meanings of visual words through language.
- 🌈 Language affects how we categorize colors and spatial relationships.
- 💡 Language provides cognitive advantages in development.
- 🔬 Empirical research supports Landau's theories.
Cronología
- 00:00:00 - 00:05:00
Susan Goldin-Meadow introduces Barbara Landau, highlighting her academic background and significant contributions to research on spatial cognition and language development in children, including those with disabilities.
- 00:05:00 - 00:10:00
Barbara Landau expresses gratitude for the William James Award and introduces her talk on whether language changes thought, outlining two perspectives on the interaction between language and thought.
- 00:10:00 - 00:15:00
Landau discusses the blind men and the elephant analogy to illustrate how different perspectives lead to different conclusions about whether language affects thought, presenting two opposing views in the field.
- 00:15:00 - 00:20:00
She reflects on her early research interests, particularly regarding the development of language in congenitally blind children, and how they acquire meanings for words despite lacking sensory experiences.
- 00:20:00 - 00:25:00
Landau introduces the Worfian hypothesis, which suggests that the language one speaks influences their thought processes, and discusses its implications in popular literature and research.
- 00:25:00 - 00:30:00
She presents two versions of the Worfian hypothesis: one suggesting that language reshapes non-linguistic thought, and the other proposing that having a language radically transforms thought, particularly in spatial reasoning.
- 00:30:00 - 00:35:00
Landau critiques the first version, questioning the clarity of what changes in perceptual discrimination, and discusses the second version's reliance on language as a necessary component for cognitive transformation.
- 00:35:00 - 00:40:00
She proposes a third hypothesis, the recoating hypothesis, suggesting that language provides a formalism that aids cognition without altering non-linguistic representations, illustrated through experiments on color and location binding in children.
- 00:40:00 - 00:45:00
Landau shares experimental findings showing that language can enhance cognitive performance in tasks involving color and location, emphasizing the momentary effects of linguistic input on cognition.
- 00:45:00 - 00:53:37
In conclusion, Landau asserts that while language does not fundamentally change non-linguistic thought, it provides significant advantages in cognitive processes, particularly through its structural properties.
Mapa mental
Vídeo de preguntas y respuestas
Who is Barbara Landau?
Barbara Landau is a professor and researcher known for her work on language and cognition, particularly in children.
What is the main focus of Barbara Landau's research?
Her research focuses on how language influences thought and cognitive development, especially in children with disabilities.
What are the two main hypotheses regarding language and thought?
One hypothesis suggests that language shapes non-linguistic thought, while the other proposes that having language transforms thought processes.
What is the recoating hypothesis?
The recoating hypothesis suggests that language provides a formalism that enhances cognitive abilities without changing underlying non-linguistic representations.
How does language affect cognition according to Landau?
Language has momentary effects on cognition, providing advantages through its structure and formalism.
What role does syntax play in language and thought?
Syntax is crucial as it helps to specify relationships and alters the prominence of entities in thought.
What is the significance of Landau's work on blind children?
Her work shows that blind children can learn the meanings of visual words despite lacking sensory input, highlighting the role of language structure.
What are some examples of how language influences thought?
Language can influence how we categorize colors, understand spatial relationships, and even infer mental states.
What is the relationship between language and cognitive development in children?
Language plays a significant role in cognitive development, providing tools for children to enhance their understanding of the world.
What is the importance of empirical research in Landau's work?
Empirical research provides evidence for her theories on the interaction between language and thought, supporting her hypotheses.
Ver más resúmenes de vídeos
- 00:00:04okay good morning everyone my name is
- 00:00:08susan goldin-meadow and it is a really
- 00:00:10tremendous honor and privilege for me to
- 00:00:12be able to introduce barbara land out to
- 00:00:15you barbara did her doctoral studies at
- 00:00:17the university of pennsylvania where she
- 00:00:19worked with lila glieben who was also my
- 00:00:21advisor a number of years earlier which
- 00:00:23makes barbara my academic sister which
- 00:00:25is really terrific since i don't have
- 00:00:26any sisters so i love it
- 00:00:28Barbara has been bicoastal in her jobs
- 00:00:31having successfully skirted the Midwest
- 00:00:33her first job was on the East Coast at
- 00:00:36Columbia University she then went to the
- 00:00:38University of California at Irvine and
- 00:00:40finally came back to the east to the
- 00:00:41University of Delaware where she is now
- 00:00:43the Johns Hopkins universe where she is
- 00:00:46now the dick and Lydia Todd professor
- 00:00:49and director of the science of learning
- 00:00:50Institute so Barbara has done seminal
- 00:00:53research on how experiential variation
- 00:00:56and genetic variation interact with
- 00:00:58developmental process to promote or to
- 00:01:01limit the development of spatial
- 00:01:03cognition and language she not only
- 00:01:05studies typically developing children
- 00:01:07but also can generally blind children
- 00:01:09and children with Williams syndrome in
- 00:01:11our early groundbreaking studies Barbara
- 00:01:14showed that congenitally blind infants
- 00:01:16acquire language just like sighted
- 00:01:18children including words such as look
- 00:01:22and see which you might not think blind
- 00:01:23children would be able to understand and
- 00:01:25Barbara show that they did she also
- 00:01:27showed that blind children can navigate
- 00:01:29untraveled routes in a novel spatial
- 00:01:32environment making it clear that they
- 00:01:34have intact geometric representations
- 00:01:36and spatial reasoning but Barbara's not
- 00:01:39only made empirical contributions she's
- 00:01:41also proposed an important theory
- 00:01:43involving the differential maturational
- 00:01:45rates of dorsal and ventral streams in
- 00:01:47the brain that explains how spatial
- 00:01:49representation and language can go awry
- 00:01:52in development Barbara's theory provides
- 00:01:55a novel explanation of the abilities and
- 00:01:57disabilities of in Williams syndrome a
- 00:01:59rare genetic disorder with an unusual
- 00:02:02pattern of cognition that spares
- 00:02:04language but profoundly impairs spatial
- 00:02:07cognition so as you know Barbara is
- 00:02:10receiving the William James Award for
- 00:02:12distinguished achievements in
- 00:02:13Psychological Science today and you will
- 00:02:15soon see why and she tells us about
- 00:02:18why how language changes thought in her
- 00:02:20talk Thank You Susan
- 00:02:28very much for that kind introduction and
- 00:02:30it's a pleasure to be here and of course
- 00:02:32I'm very grateful to and honored to have
- 00:02:35received the William James fellow award
- 00:02:39Susan mentioned that I do work on a
- 00:02:41typically developing children I'm going
- 00:02:43to talk about none of that today just
- 00:02:45for those of you who are wondering it is
- 00:02:46all about typically developing children
- 00:02:48and the question that I'm going to ask
- 00:02:50is whether language changes thought so
- 00:02:55in general the field has two ways of
- 00:02:58thinking about the language and thought
- 00:02:59interaction there are two directions
- 00:03:01that one can consider one direction is
- 00:03:04thinking about whether thought
- 00:03:06pre-linguistic thought serves as the the
- 00:03:10the the support for language learning
- 00:03:13and we have very distinguished
- 00:03:16philosophers and psychologists jerry
- 00:03:18fodor Rene Byers Aldous balki who have
- 00:03:20done tremendous work on understanding
- 00:03:23what the pre linguistic logic could be
- 00:03:26that could support language learning the
- 00:03:29other way of thinking about the language
- 00:03:30thought interaction is that language
- 00:03:34effects thought and of course that's
- 00:03:36what I'm going to focus on today within
- 00:03:39this issue there again two different
- 00:03:41ways of thinking about whether or not
- 00:03:43language effects thought so one way is
- 00:03:46to think about whether learning a
- 00:03:47particular language changes non
- 00:03:50linguistic thought so we have more than
- 00:03:525,000 languages today across the world
- 00:03:55and the question is if you learn
- 00:03:56language a versus language B does that
- 00:03:58change your non linguistic
- 00:04:00representations or the way that you
- 00:04:01think non linguistically the other way
- 00:04:04of thinking about this question whether
- 00:04:05language affects thought is to think
- 00:04:07about what happens if you have a
- 00:04:09language that is our newest species that
- 00:04:12has a language and so here we can
- 00:04:13compare human babies and young children
- 00:04:16who of course do acquire language very
- 00:04:18effortlessly early in life and the
- 00:04:21language has very deep properties as
- 00:04:23compared to let's say chicks or mice or
- 00:04:26rats who will never have a language they
- 00:04:28just simply don't have the capacities
- 00:04:29languages species specific
- 00:04:31so there are two different kinds of
- 00:04:33questions whether learning a particular
- 00:04:35language a versus B changes non
- 00:04:36linguistic thought and then there's
- 00:04:38another question whether having any
- 00:04:40language at all changes non linguistic
- 00:04:42thought so does language change thought
- 00:04:45and if so how well I'm going to argue
- 00:04:49that the answer of this question depends
- 00:04:51in part on what part of the problem you
- 00:04:54happen to be looking at and here we have
- 00:04:57the the blind man with the elephant
- 00:05:00blind men exploring the elephant and for
- 00:05:02those of you who aren't familiar with
- 00:05:03this with this sort of cartoonish
- 00:05:08instantiation of the question every
- 00:05:11person blindfolded person is exploring
- 00:05:14the elephant and the question is what is
- 00:05:16an elephant and of course depending on
- 00:05:17where you look depending on where you're
- 00:05:19feeling you have a different answer so
- 00:05:21for this person the elephant might be
- 00:05:24something that's long and thin for this
- 00:05:25person the elephant might be something
- 00:05:27that's broad and wrinkly and so forth so
- 00:05:29it depends on where you're looking
- 00:05:31locally whether or not what you think
- 00:05:34the elephant is similarly if you're
- 00:05:37asking the question does language change
- 00:05:38thought the field in general depending
- 00:05:40on where they have looked have come up
- 00:05:42with two different answers very distinct
- 00:05:44ones and those are definitely yes it
- 00:05:46does and definitely not no it doesn't so
- 00:05:50that's what we're going to try to deal
- 00:05:51with today and what I'm gonna do in the
- 00:05:55talk is to review it basically two
- 00:05:57prominent hypotheses that I sort of
- 00:05:59think capture the way that people think
- 00:06:01about the language thought interaction
- 00:06:03whether language affects thought and I'm
- 00:06:05gonna offer a third hypothesis which I
- 00:06:07think accounts for the existing data as
- 00:06:08well as much other data but first I want
- 00:06:11to go back in history a little bit and
- 00:06:13give you a little bit of background with
- 00:06:14respect to my own interests in language
- 00:06:17and thought and Susan already covered
- 00:06:18some of this the early probe for me
- 00:06:21about the language thought relationship
- 00:06:23had to do with the growth and
- 00:06:25development of a child who was born
- 00:06:27blind the child who's who's blind from
- 00:06:29birth and my question at the time was
- 00:06:32what is the role of sensory and
- 00:06:33perceptual experience and how does it
- 00:06:35interact with what I'm gonna call mind
- 00:06:37driven aspects of our knowledge that is
- 00:06:39to say some of our information comes
- 00:06:40from exploring the world either visually
- 00:06:43or haptically if we're blind
- 00:06:44but some information and some knowledge
- 00:06:46comes specifically from the mind itself
- 00:06:49and this work which was done in
- 00:06:52conjunction with Lila Blackman who was
- 00:06:53my advisor at the time of my PhD drove
- 00:06:58us to some some very substantial
- 00:07:00conclusions which I'll just detail very
- 00:07:02very quickly that really do emphasize
- 00:07:05the importance of having a mind a human
- 00:07:08mind in exploring and learning things
- 00:07:10about the world when I began my thinking
- 00:07:15about this it actually was as Susan said
- 00:07:18I arrived just after she had done her
- 00:07:20dissertation and she and heidi Feldman
- 00:07:22and Lila Lightman had at that time
- 00:07:24discovered that children born
- 00:07:25congenitally deaf could nevertheless
- 00:07:27construct a gestural language without
- 00:07:29linguistic input and I came into
- 00:07:31graduate school and I said well that's
- 00:07:33fine but those people those those those
- 00:07:36deaf individuals they had the means to
- 00:07:38interpret the world because they could
- 00:07:39see and so I proposed that we should
- 00:07:42study the blind child - as an obvious
- 00:07:44compliment a child who had full
- 00:07:46linguistic input but much reduced
- 00:07:49experience through vision and the study
- 00:07:52took two parts as I mentioned yesterday
- 00:07:54or the day before at the opening
- 00:07:57comments my advisers encouraged me to
- 00:07:59read very broadly and one of the people
- 00:08:02that I read was William James and I
- 00:08:03couldn't help but go back and remember
- 00:08:06and and find the paragraphs that he
- 00:08:09wrote that I found so meaningful when I
- 00:08:10was first thinking about the blind child
- 00:08:12and whether or not they could even
- 00:08:14develop a sense of space and William
- 00:08:16James said the seeing baby's eyes taken
- 00:08:18the whole room at once and
- 00:08:19discrimination must arise in him before
- 00:08:22single objects are visually discerned
- 00:08:24the blind child on the contrary must
- 00:08:26formed his mental image of the room by
- 00:08:28the addition piece to piece of parts
- 00:08:30which he learns to know successively so
- 00:08:33that implies the input is very very
- 00:08:35different for a blind child than a
- 00:08:36sighted child and yet what this Balki
- 00:08:39and Henry Lightman and I found in
- 00:08:41looking at spatial knowledge was that
- 00:08:43the blind child even a two-year-old
- 00:08:44blind child and a cobbler and a sighted
- 00:08:48child as well
- 00:08:48they can both construct the same
- 00:08:50geometric representations of layouts in
- 00:08:52the world despite these vast differences
- 00:08:54in experience the second part to my
- 00:08:57questions at the time
- 00:08:58to do with language in the blind and
- 00:08:59here Lila and I were thinking about well
- 00:09:01what would happen if you didn't have a
- 00:09:03means to interpret the world and of
- 00:09:05course we relied on John Locke who said
- 00:09:07many things about the blind he was a
- 00:09:09very diehard empiricist and the quote
- 00:09:12that best represents his view is I think
- 00:09:14if it will be granted easily that if a
- 00:09:17child were kept in a place where he
- 00:09:18never saw any other but black and white
- 00:09:20till he were a man he would have no more
- 00:09:22ideas of scarlet or green than he that
- 00:09:25from his childhood never tasted an
- 00:09:27oyster or a pineapple has of those
- 00:09:29relishes that is to say if you don't
- 00:09:31have sensory input to feed a lexical
- 00:09:35concept a concept or a word for that
- 00:09:37concept then you will never develop them
- 00:09:39but Lila and I found that the blind
- 00:09:42child actually does learn the meanings
- 00:09:44of words like look and see and color
- 00:09:46terms despite the obvious lack of any
- 00:09:49sensory or perceptual info for these
- 00:09:51particular words and this is this panel
- 00:09:54actually shows a picture of our our main
- 00:09:58participant whose name was Kelly
- 00:10:00responding to our command to look up and
- 00:10:03so Kelly's interpretation over the word
- 00:10:05look was to discover things with your
- 00:10:07hands and so when we said look up to the
- 00:10:10sighted child the sighted child of
- 00:10:11course goes like this even by the way if
- 00:10:13they're blindfolded but if a blind child
- 00:10:16if this blind child is told to look up
- 00:10:18they go like this reaching in order to
- 00:10:20explore whatever is above so Kelly
- 00:10:23definitely the sighted the blind child
- 00:10:25did definitely develop really quite
- 00:10:28detailed representations of those visual
- 00:10:30words so that's my background and that's
- 00:10:33how I came to think about the broader
- 00:10:34question of whether language affects
- 00:10:36thought in particular for Kelly our
- 00:10:39theory was that the way that she learned
- 00:10:40the meanings for these words look and
- 00:10:42see as well as color terms was actually
- 00:10:44through the structure of the language
- 00:10:46itself so I will be saying more about
- 00:10:48the structure of the language itself and
- 00:10:50its power throughout the talk but that
- 00:10:52was our hypothesis at the time and
- 00:10:55that's the question does language affect
- 00:10:57thought in some way that's continued to
- 00:10:59occupy me for the rest of my career
- 00:11:02there's language of that thought and if
- 00:11:04so how so now I want to fast-forward to
- 00:11:08sort of now
- 00:11:10and say what the Hubertus take on the
- 00:11:13language affecting thought hypothesis it
- 00:11:15question is probably everybody in this
- 00:11:19room knows about the Worf ian hypothesis
- 00:11:21the idea is the language that you speak
- 00:11:23the particular language that you speak
- 00:11:24causes you to let's say view the world
- 00:11:27think about the world in a very
- 00:11:29different way if you're like learning
- 00:11:31language a versus B and this was a quote
- 00:11:33that was published in the New York Times
- 00:11:35Magazine section from 2010 it's a part
- 00:11:38of a review of a book by a person named
- 00:11:40guy deutscher who very much endorsed the
- 00:11:43wharf Ian view and I just want to read a
- 00:11:45bit of this to you so that you'll get
- 00:11:47the flavor of the wharf Ian hypothesis
- 00:11:49when your language routinely obliges you
- 00:11:52to specify certain types of information
- 00:11:53it forces you to be attentive to certain
- 00:11:56details in the world and to certain
- 00:11:58aspects of experience that speakers of
- 00:11:59other languages may not be required to
- 00:12:01think about all the time emphasis mine
- 00:12:04and the next and since such habits of
- 00:12:06speech are cultivated from the earliest
- 00:12:08stage it's only natural that they settle
- 00:12:10into habits of mind that go beyond
- 00:12:12language itself affecting your
- 00:12:14experiences your perceptions your
- 00:12:17associations your feelings your memories
- 00:12:20and your orientation in the world notice
- 00:12:23that is a enormous ly strong hypothesis
- 00:12:25that it affects everything you could
- 00:12:27possibly imagine now I want to say that
- 00:12:30this hypothesis has actually led to
- 00:12:32quite strong claims especially in the
- 00:12:35popular literature so I'm going to use
- 00:12:37as an illustration a TED talk that was
- 00:12:41given in 2013 by a behavioral economist
- 00:12:43named Keith Chen from UCLA and Chen's
- 00:12:46thesis was that speaking a language that
- 00:12:49doesn't obligatory that is by necessity
- 00:12:51Marc future tense such as Chinese which
- 00:12:54was his native language strongly
- 00:12:56correlates with higher savings compared
- 00:12:59to speakers of languages that do not
- 00:13:01that do obligatory Li Marc tense as in
- 00:13:04English this is a TED talk widely
- 00:13:07distributed and I think anybody who
- 00:13:10pauses for just a minute can realize
- 00:13:11that that's actually sort of a puzzling
- 00:13:13hypothesis and probably what happened is
- 00:13:15that they discovered what the savings
- 00:13:17rates were and then went back and
- 00:13:18conjectured from what the patterns of
- 00:13:20the language were but that's obviously
- 00:13:22just adjusts
- 00:13:23story and so I want to say to you that
- 00:13:26when you read things on popular
- 00:13:27literature about language affecting
- 00:13:29thought you must be very careful because
- 00:13:31it's not a good thing to be promulgating
- 00:13:34really broad and deep hypotheses on the
- 00:13:37basis of really quite flimsy data or
- 00:13:39questionable data in any case we're
- 00:13:42lucky that most of the psyche the
- 00:13:43psychological science that's done is not
- 00:13:45of this ilk and so now I want to go to
- 00:13:48the two versions of the the of language
- 00:13:52and thought that are normally talked
- 00:13:54about that are sort of part of the
- 00:13:55conversation in the literature the first
- 00:13:57I'm going to call the classic Worf Ian's
- 00:13:58hypothesis and this is the idea that
- 00:14:00language reshapes non linguistic thought
- 00:14:03whatever your non linguistic thought is
- 00:14:05if you have language a versus language
- 00:14:07be it will reshape it and one of the
- 00:14:09examples that I want to give you is
- 00:14:10about color terms which has been a
- 00:14:12testbed for a very long time for the
- 00:14:14Worf Ian hypothesis because languages
- 00:14:16tend to cut up the the space of colors
- 00:14:19in very different ways across different
- 00:14:21languages it's not unconstrained but
- 00:14:22there are differences and so the idea
- 00:14:24here is that learning a particular
- 00:14:26language language a versus language B
- 00:14:28shapes non linguistic categories in
- 00:14:31accord with the native language
- 00:14:33distinction this has been done not only
- 00:14:36in English the study people have used
- 00:14:38English but they've also used a variety
- 00:14:39of other languages to test the
- 00:14:41hypothesis that if you have a
- 00:14:42distinction that you make in one
- 00:14:45language that's a different distinction
- 00:14:46in another language that is across as
- 00:14:48different boundaries you will have
- 00:14:49different perceptual discrimination and
- 00:14:51I'll tell you about this experiment in
- 00:14:53just a second but basically the idea
- 00:14:55here is we have four shades that range
- 00:14:57from sort of pretty much green to pretty
- 00:14:59much blue these two patches these are by
- 00:15:03the way all of equal values physically
- 00:15:06psychologically they're somewhat
- 00:15:09different but in terms of naming they're
- 00:15:11very different so it tends to be the
- 00:15:13case that if you have hues that look
- 00:15:14roughly like this people tend to call
- 00:15:17them green and if you have hues that
- 00:15:18look roughly like this people tend to
- 00:15:21call them blue and so what happens is
- 00:15:23that this is called a between category
- 00:15:26distinction these are the greens and
- 00:15:28these are the blues and this is called a
- 00:15:30within category distinction there are
- 00:15:32two different types of greens or two
- 00:15:34different types of blues and so the idea
- 00:15:36here
- 00:15:37that if your language shapes your
- 00:15:39perceptual discrimination as Gilbert at
- 00:15:41all have hypothesized then what you
- 00:15:44should see is that people who speak a
- 00:15:45language that has greens and blues here
- 00:15:48should show a bigger difference between
- 00:15:50these two shades than people who speak a
- 00:15:52different language now I'm not going to
- 00:15:53tell you about the cross linguistic data
- 00:15:55but I'll just tell you the snippet of
- 00:15:57the English data that will suffice for
- 00:15:59us I want to say though that the
- 00:16:00conclusion of this study by Gilbert at
- 00:16:03all in 2006 was that language may affect
- 00:16:05perceptual discrimination words are
- 00:16:08important here okay it's about
- 00:16:09discrimination it's not about anything
- 00:16:11else it's perceptual discrimination so
- 00:16:14here's the experiment basically that
- 00:16:16they gave people patches of colors in
- 00:16:21which there was one outlier so in this
- 00:16:23case the outlier is the thing that looks
- 00:16:26that's called blue by English speakers
- 00:16:27and the other elements are all in this
- 00:16:30category over here so that's a between
- 00:16:32category distinction and people were
- 00:16:35asked to fixate here and then the panel
- 00:16:37was brought on with all of these colors
- 00:16:38and all they had to do was identify the
- 00:16:40outlier and they're going to ask whether
- 00:16:42or not it's quicker for people to
- 00:16:44identify the outlier if it is represents
- 00:16:47a between linguistic category
- 00:16:49distinction compared to if it's a within
- 00:16:53distinction category and they went a
- 00:16:55little bit farther they said actually
- 00:16:56it's because of the way the brain works
- 00:16:58and so what they did was they actually
- 00:16:59presented to the right visual field in
- 00:17:01which case it projects to the left
- 00:17:03hemisphere or they projected to the left
- 00:17:05hemisphere field in which case of
- 00:17:07projects to the right hemisphere and
- 00:17:09what they found was indeed the between
- 00:17:12categories distinctions were faster than
- 00:17:14the within category as you would expect
- 00:17:16if language had affected something about
- 00:17:19your decision but it was modulated by
- 00:17:21which visual field that was presented in
- 00:17:23and it was the distinction was that
- 00:17:25there was a faster reaction time for
- 00:17:27these between category judgments only in
- 00:17:30the right visual field that is
- 00:17:31projecting the left hemisphere so and in
- 00:17:34addition they presented some data that
- 00:17:36suggested that the pattern was disrupted
- 00:17:38if there was a verbal interference at
- 00:17:39the same time that if you were ready to
- 00:17:41do something with language that effect
- 00:17:42would go away so it seemed to them it
- 00:17:44was a linguistic effect but their
- 00:17:45interpretation was very strong and again
- 00:17:47words matter language may affect
- 00:17:50perceptual disk
- 00:17:51nation okay I think that's gone too far
- 00:17:53I'll say why in just a minute but that's
- 00:17:58an example of version one the particular
- 00:18:00language you speak affects the way that
- 00:18:03you nan your non linguistic
- 00:18:05representations the second version that
- 00:18:08I think is actually more interesting in
- 00:18:10a way is that language having a language
- 00:18:12causes a radical transformation of
- 00:18:14thought and the example I want to give
- 00:18:16here is spatial reorientation and for
- 00:18:18those of you who are not spatial
- 00:18:20reorientation aficionado so I'll go
- 00:18:21through this briefly and try to make it
- 00:18:25clear but it turns out there's there's a
- 00:18:28very interesting phenomenon within the
- 00:18:30literature on spatial reorientation
- 00:18:31really by reorientation we mean if you
- 00:18:34get disoriented in a space and then you
- 00:18:36must reorient yourself what information
- 00:18:38do you use and some of this work
- 00:18:40actually the seminal work was done by
- 00:18:42Randy gal Estelle and Ken Chang back in
- 00:18:45the early 80s in which they found that
- 00:18:46if you put a rat into a chamber that is
- 00:18:49roughly rectangular and you bait it at
- 00:18:52one corner and then you take the rat out
- 00:18:55and disorient the rat and put the rat
- 00:18:58back in in the middle they actually make
- 00:19:00a very peculiar error sometimes they go
- 00:19:03to the baited side the baited corner but
- 00:19:05sometimes they go to the corner that's
- 00:19:07exactly geometrically opposite that
- 00:19:09baited corner and the question is why
- 00:19:11and Chang and gallo still proposed that
- 00:19:13actually they were ignoring all of the
- 00:19:16salient visual cues I mean rats are not
- 00:19:18visual anyway but in you know smell and
- 00:19:20so forth they you were ignoring those
- 00:19:22surface cues in order to focus on what
- 00:19:25we call the geometry of the space so in
- 00:19:28geometric terms this corner with the
- 00:19:30long wall let's say to the left and the
- 00:19:32short wall to the right is actually
- 00:19:34geometrically equivalent in to this
- 00:19:37corner so the rats were essentially
- 00:19:38recording the geometry of the space and
- 00:19:41were essentially ignoring the surface
- 00:19:43cues now Linda Hermer and Liz spelke
- 00:19:45then adopted the method of disorienting
- 00:19:48animals in these chambers and they
- 00:19:51adopted it for young children so they
- 00:19:53took 18 and 24 month olds and they also
- 00:19:55put them in a rectangular chamber like
- 00:19:57this and actually it was sort of
- 00:19:59originally just a black chamber with no
- 00:20:02landmarks at all
- 00:20:03and they showed them a toy that was
- 00:20:06hidden in this corner and then they
- 00:20:08disoriented them close their eyes and
- 00:20:10turned them around and they found that
- 00:20:1218 to 24 month olds and actually just
- 00:20:16about every age that's been tested if
- 00:20:18there are no surface cue so if the what
- 00:20:20if the if the enclosure is all black and
- 00:20:23there are no landmarks at all children
- 00:20:26and adults make the same geometric error
- 00:20:28as rats okay
- 00:20:30the peculiar thing in addition to that
- 00:20:32was the very youngest children really
- 00:20:34seem to act just like rats
- 00:20:36and so they too ignored the surface cues
- 00:20:39they ignored the landmarks and basically
- 00:20:41went with geometry later on these
- 00:20:44children when you get to be you know
- 00:20:46four or five or six years old and
- 00:20:47certainly adults if you have the
- 00:20:49landmarks as well then it seems you
- 00:20:51combine the information and so then you
- 00:20:53get it exactly right and so you can find
- 00:20:55the object in the right corner okay so
- 00:20:58the proposal the reason this is
- 00:21:00important is because they had a proposal
- 00:21:02about why that changes why does the yeah
- 00:21:04what is a young child a toddler act just
- 00:21:06like a rat but a four to six year old
- 00:21:08now acts like an adult that is they can
- 00:21:10combine the geometry with the landmarks
- 00:21:12and their answer was this the
- 00:21:14combinatorial properties of the language
- 00:21:16faculty serve to represent relationships
- 00:21:19the child's nonverbal systems cannot
- 00:21:21capture because of encapsulation of
- 00:21:23those systems that is to say they're
- 00:21:24represented in cept in separate places
- 00:21:26the geometry and the landmarks are
- 00:21:28separate but if you have language and
- 00:21:30you can represent it linguistically then
- 00:21:32that's a huge advantage okay so that's
- 00:21:35their hypothesis that it's language
- 00:21:37itself having a language that is causing
- 00:21:39the difference between the rat and the
- 00:21:41human this hypothesis which I think has
- 00:21:44some real merit has actually been taken
- 00:21:47very seriously in the philosophic Asafa
- 00:21:50chol and linguistic literature those
- 00:21:52bodies people really have admired this
- 00:21:54so the philosopher Carruthers says
- 00:21:56Hermer vasquez it all provides strong
- 00:21:58evidence that the integration of
- 00:22:00geometry with other sorts of information
- 00:22:02depends on natural language okay and
- 00:22:05Berwick and Chomsky recently have asked
- 00:22:08why do humans have language at all
- 00:22:09because language is the lingua franca
- 00:22:12the binds together the different
- 00:22:14representations from geometric
- 00:22:16non-geometric modules just like an inner
- 00:22:18mental tool should so people seem to
- 00:22:23admire this hypothesis in the idea that
- 00:22:26having a language does something
- 00:22:27extraordinary for you these versions
- 00:22:30both version 1 and version 2 are
- 00:22:32interesting but they really have
- 00:22:33critical weaknesses so for example the
- 00:22:35language shapes thought hypothesis
- 00:22:37version 1 it's really unclear what's
- 00:22:39changing was it really a change in
- 00:22:41perceptual discrimination I really doubt
- 00:22:43it could it have been memory it's
- 00:22:45possible it's possible that it's just a
- 00:22:46linguistic effect that is to say you're
- 00:22:49coding it linguistically and therefore
- 00:22:51you're showing the effects of language
- 00:22:53secondly very often what's called a non
- 00:22:56linguistic task in this field actually
- 00:22:58is is permeated with language so it
- 00:23:02actually is a linguistic task and so you
- 00:23:04have to be very careful about that and
- 00:23:05finally the other weakness is that this
- 00:23:07hypothesis requires that you use a
- 00:23:10language habitually and we don't know
- 00:23:12what the length of time is is that over
- 00:23:14the lifetime and then you get to the
- 00:23:16point that your savings changes because
- 00:23:18you have a language that marks the
- 00:23:21future tense unclear the second
- 00:23:23hypothesis the version 2 that radically
- 00:23:25linked language radically transforms
- 00:23:27thought it the strong version of this
- 00:23:29that is necessary language is necessary
- 00:23:31it's already ruled out by evidence from
- 00:23:33animals so animals do combine it's not
- 00:23:35that like they act like the Chang and
- 00:23:37galas tell rats ignoring the surface
- 00:23:39cues at all so we don't need language to
- 00:23:41do this but it's also a problem that the
- 00:23:43mechanism seems completely unspecified
- 00:23:45and I think this is a huge problem in
- 00:23:47the field which I hope to provide part
- 00:23:49of a solution to this morning okay I
- 00:23:51want to move on to the third hypothesis
- 00:23:53that I'm going to propose and I'm going
- 00:23:55to call it the recoating hypothesis in
- 00:23:58the past I've called it at the momentary
- 00:23:59hypothesis for anybody who has an
- 00:24:01inkling here of the past work but we'll
- 00:24:05call it the recoating hypothesis right
- 00:24:06now and I'm going to spell out what I
- 00:24:08mean by this and show you an example of
- 00:24:10how it could work with some experiments
- 00:24:12so the idea here is that language
- 00:24:14provides a rich formalism it's not what
- 00:24:17it means to have a language is to have a
- 00:24:19formalism that really matters when
- 00:24:21you're encoding your experience and that
- 00:24:24itself having that formalism converts a
- 00:24:26huge advantage for cognition
- 00:24:28and the case study I'm going to be
- 00:24:30dealing with is a very simple case of
- 00:24:33binding color and location storing
- 00:24:35colors in their locations these are the
- 00:24:38references the for the the papers in
- 00:24:40which the the empirical data has been
- 00:24:42reported this person bunchy de Celine
- 00:24:44was my graduate student at the time that
- 00:24:45we started this she's done she did most
- 00:24:47of the experimental work with me and so
- 00:24:50let me spell it out the recoding
- 00:24:54hypothesis says language recode x' what
- 00:24:56we see it provides a rich
- 00:24:57representational formalism for recoding
- 00:25:00what we see it includes both semantics
- 00:25:03the meanings of words and it also
- 00:25:04couldn't include syntax the way that we
- 00:25:06combine words the benefit of recoding
- 00:25:09occurs right in the moment of doing the
- 00:25:13task its role changes over development
- 00:25:15and its enormous ly powerful it can lead
- 00:25:17to cognitive enrichment but it does not
- 00:25:20change underlying non linguistic
- 00:25:21representations so it's not effective on
- 00:25:23perceptual discrimination so for some
- 00:25:26background i'm gonna tell you feature
- 00:25:28conjunctions are hard so a lot of people
- 00:25:31are here if you have any background in
- 00:25:32vision you'll recognize these effects
- 00:25:34very quickly visual search in adults
- 00:25:37shows that if you're searching for a
- 00:25:39feature for a color a particular color
- 00:25:41in an array or a particular shape is
- 00:25:43really fast and it's called parallel but
- 00:25:46a conjunction search if you're searching
- 00:25:48for a color and shape it's very slow and
- 00:25:50it tends to be serial you have to look
- 00:25:52one by one it requires attention and
- 00:25:54this was one of the founding research
- 00:25:57findings from Anne Treisman back in 1984
- 00:26:01so I'm gonna give you an example and I
- 00:26:03want a little audience participation
- 00:26:04here I want you to look for the tar this
- 00:26:07target okay in the next display here I
- 00:26:09go
- 00:26:10how many found it everybody found it it
- 00:26:13popped out it was immediate it was
- 00:26:15totally easy now let's do the same
- 00:26:17exercise look for the target the same
- 00:26:19target okay raise your hand when you see
- 00:26:25it and those of you who haven't found it
- 00:26:26yet it's okay you will find it I promise
- 00:26:29you okay but you do need to go serially
- 00:26:32one by one so that just shows you the
- 00:26:34binding of the color and the location or
- 00:26:36the color and the shape is really really
- 00:26:38hard for the visual system to do that's
- 00:26:40think that's a weakness of on the part
- 00:26:42of the visual system in some way the
- 00:26:44general point is that visual search
- 00:26:46conjunction search are hard turns out
- 00:26:49patients with vision disorder disorders
- 00:26:51due to brain damage show deficits in
- 00:26:53this particular kind of search it also
- 00:26:55turns out that memory for color and
- 00:26:57location is really poor in typically
- 00:26:59developing four to six year olds and in
- 00:27:01adolescents with Williams syndrome and I
- 00:27:03have a picture of a child one of our
- 00:27:05earlier participants with Williams
- 00:27:06syndrome be good to remind me this is
- 00:27:08actually how we sort of stumbled on this
- 00:27:10this this method in this particular case
- 00:27:13which turns out to be very important we
- 00:27:15noticed that these children had a
- 00:27:17tremendous difficulty representing the
- 00:27:19color and location of something as
- 00:27:21simple as this stimulus and I'm going to
- 00:27:23show you a lot of data from typically
- 00:27:24developing four-year-olds that show the
- 00:27:26same so we began to think could it be
- 00:27:28possible that this weakness in the
- 00:27:30visual system is aided by language and
- 00:27:34so I'm going to ask three questions
- 00:27:35first of all if there is change is it
- 00:27:37long short or momentary what's the
- 00:27:41nature of the linguistic representation
- 00:27:43that affects the change and how powerful
- 00:27:46is it relative to other mechanisms and
- 00:27:48these are going to be my answers it's
- 00:27:50momentary it involves lexical that is
- 00:27:52the word meanings plus the syntax it's
- 00:27:55very abstract and it's very powerful so
- 00:27:58here I'm going to show you the basic
- 00:27:59task and this is all gonna be with
- 00:28:00four-year-olds unless I so say it say
- 00:28:02something unless I tell you otherwise
- 00:28:03and here's the task you show a
- 00:28:06four-year-old child this thing and you
- 00:28:07say look at this and then they can look
- 00:28:10at it as long as they want and then it
- 00:28:12goes off for one second one second and
- 00:28:14then three options come on and we say
- 00:28:17which one is exactly the same as the one
- 00:28:19you just saw so probably everybody knows
- 00:28:22which one but if you're confused
- 00:28:23do not worry it's very common so I
- 00:28:26actually can't remember which one was in
- 00:28:28the first panel but the typical errors
- 00:28:30are between these two because of the
- 00:28:32color location one is a reflection of
- 00:28:33the other so in the task we had three
- 00:28:37different target types we had a vertical
- 00:28:38split a horizontal split and a diagonal
- 00:28:40split and for each one of these
- 00:28:43different target types for each trial
- 00:28:45the structure of the trials were just
- 00:28:48like I showed you and just sort of more
- 00:28:49abstractly if I see if they saw a
- 00:28:51vertical split they would get the same
- 00:28:52one
- 00:28:53and it's reflection and then they would
- 00:28:56get one of the other splits so either
- 00:28:57they would get a horizontal split or
- 00:28:59they would get a diagonal split so
- 00:29:00always three options when after the
- 00:29:03after the target goes off I want to make
- 00:29:06a point that our native object
- 00:29:08recognition system actually considers
- 00:29:10these things reflections to be
- 00:29:11equivalent so for us purposes of object
- 00:29:14recognition pretty much the visual
- 00:29:15system ignores this reflection so it
- 00:29:18makes sense for us to have problems
- 00:29:20identifying a target and its reflection
- 00:29:22we only need these the reflections and
- 00:29:25the distinctions we do them to
- 00:29:26distinguish for exerting letters and
- 00:29:28numbers so when you learn P's and B's
- 00:29:30there's a P and there's a and the B's
- 00:29:32and their DS those are different
- 00:29:34geometric transformations of each other
- 00:29:36and for those we need to mark the
- 00:29:38differences among across there the the
- 00:29:40items so here's the first experiment
- 00:29:43again four-year-old children when we
- 00:29:45started out we thought to ourselves
- 00:29:46let's just do a very very simple
- 00:29:48manipulation with language so the
- 00:29:52experiment is you show them this thing
- 00:29:54and either you do the know label you say
- 00:29:56look at this and then goes off for a
- 00:29:59second and then which one is exactly the
- 00:30:01same as you just saw or we try to novel
- 00:30:03label because there's an abundant
- 00:30:05literature including some that I've been
- 00:30:06involved in that shows that if you have
- 00:30:08a label associated with an object it
- 00:30:10tends to draw attention to the object at
- 00:30:12least that's the claim so we thought
- 00:30:13let's just label it so then we said in
- 00:30:15another condition completely different
- 00:30:17set of subjects look at this this is a
- 00:30:19Dax and then which one is exactly the
- 00:30:21same as you just saw and the results
- 00:30:24were these remember that random chance
- 00:30:27probability performance is 33% so
- 00:30:30everybody is above 33% these are the
- 00:30:32corrects are about 66% and 62% there's
- 00:30:35no difference between the label and a no
- 00:30:36label condition they're right around 60
- 00:30:39and what's interest interesting is the
- 00:30:41errors so the errors are primarily the
- 00:30:43reflections just as we predicted so
- 00:30:45they're the children are not hardly ever
- 00:30:48picking the other split if they have a
- 00:30:50horizontal split they're not picking a
- 00:30:51vertical split they're picking the right
- 00:30:53split but they're mixing up what the
- 00:30:55assignment of color to location is so
- 00:30:57this is what the first experiment showed
- 00:31:00us and my graphic for representing what
- 00:31:02the child must be representing is
- 00:31:04something like this they're representing
- 00:31:06the gene
- 00:31:06split but they're not sure which side
- 00:31:08the red and the green go on okay so it's
- 00:31:11there's no assignment of color to the
- 00:31:12locations then we thought okay let's
- 00:31:15just go whole hog here let's just say
- 00:31:17look at this the red is on the left or
- 00:31:20of course for other stimuli if the rig
- 00:31:22was on the right we said on the right
- 00:31:23top and bottom and so forth
- 00:31:24goes off so look at this the red is on
- 00:31:27the Left goes off for a second and the
- 00:31:30items come on which one is exactly the
- 00:31:32same as you just saw and here what we
- 00:31:34find is performance goes up to 80% from
- 00:31:3660% so it's a huge bump up these are
- 00:31:39again four-year-old children we get the
- 00:31:41same error patterns and this will be
- 00:31:43persistent throughout all of the
- 00:31:44experiments I tell you about and so my
- 00:31:47graphic for representing what the child
- 00:31:49knows is that now they're actually doing
- 00:31:51much better they represent not only the
- 00:31:52split but the color location assignments
- 00:31:55now we begin to think well okay we have
- 00:31:58to begin thinking about this what is
- 00:31:59language doing here we told them left
- 00:32:01and right
- 00:32:01so we say look at this let's see where
- 00:32:04the red is and then we flashed the whole
- 00:32:06or just the red part and so just to give
- 00:32:09you an example of a flashing one we show
- 00:32:11them look at this look where the rate is
- 00:32:14and then show me the one that's exactly
- 00:32:17the same and we did that with the whole
- 00:32:19as well we did flashing with the halt in
- 00:32:21the experiment three shows back down to
- 00:32:2360% so that the salience method doesn't
- 00:32:26do the trick at all and the
- 00:32:28representation is still a vertical split
- 00:32:30with nothing else follow-ups growing and
- 00:32:33pointing so our vision science
- 00:32:35colleagues said to us don't make it
- 00:32:37flash on and off make it grow because
- 00:32:39flashing on and off is like it's absent
- 00:32:40so we said fine we'll make it grow so we
- 00:32:42made it grow in one experiment I'll show
- 00:32:44you what it looks like we also had
- 00:32:46children in a separate experiment point
- 00:32:47because we're good developmental
- 00:32:49psychologists and we thought surely
- 00:32:50pointing to the red will work so here's
- 00:32:53what it looks like to grow the red side
- 00:32:56look at this look at the red and then
- 00:33:00which one is exactly the same as you've
- 00:33:03just seen and here's what happens all at
- 00:33:06the same so this is flashing this is
- 00:33:08growing and this is pointing that is to
- 00:33:10say we're back down at 60% those have
- 00:33:13absolutely no effect okay
- 00:33:15so now we think okay we said left maybe
- 00:33:18we should just give any kind
- 00:33:19spatial term let's try the red is
- 00:33:21touching the green and we used connected
- 00:33:23to in adjacent to which one is exactly
- 00:33:26the same as you sis just saw and the
- 00:33:28representation is again no better than
- 00:33:31we saw before it doesn't assign color
- 00:33:33location and here are the first four
- 00:33:34experiments that I just told you about
- 00:33:36hole or no label is 60% top bottom right
- 00:33:40and left is around 80 percent flashing
- 00:33:42pointing growing around 60% and touching
- 00:33:44which is neutral back down to 60% so the
- 00:33:47only thing that works is the red is the
- 00:33:49left of the green nothing else works
- 00:33:51nothing else that any vision scientists
- 00:33:53are developmentalist would have
- 00:33:54predicted so what is going on
- 00:33:56what does language do I'm going to start
- 00:33:58by trying to convince you that these are
- 00:34:00changes that occur right in the moment
- 00:34:02of the task for the child so what we did
- 00:34:04after the left in the left/right
- 00:34:07experiment was we also did a post task
- 00:34:10and so after the child was done with the
- 00:34:13experiment we ran a production study and
- 00:34:14a comprehension study to see whether or
- 00:34:16not children who knew the meanings of
- 00:34:19left and right that is this is on my
- 00:34:20left and this is on my right did better
- 00:34:22because they had a long-term
- 00:34:24representation of that term which was
- 00:34:26brought to bear on the task then
- 00:34:28children who didn't know it so in the
- 00:34:30production task we had a smiley face we
- 00:34:32presented at one of four locations and
- 00:34:34we said where's the smiley face it is
- 00:34:35and then they say they give a spatial
- 00:34:38term in the comprehension task we said
- 00:34:41see this edge they see the square draw
- 00:34:44an X on the left right top bottom of the
- 00:34:47square and so that's a comprehension to
- 00:34:48ask and they placed an X wherever they
- 00:34:50thought here are the results which
- 00:34:52aren't too surprising top and bottom for
- 00:34:55four-year-olds at ceiling they know
- 00:34:57they're tops and bottoms but they do not
- 00:34:58know they're left's and rights okay and
- 00:35:00it varies it varies enormous ly which is
- 00:35:02a good thing for us because there are
- 00:35:03some kids who just don't know left-right
- 00:35:05at all and there's some kids who are all
- 00:35:07the way up at ceiling at 80% for these
- 00:35:09things I have to say that even the kids
- 00:35:11who don't know left and right do know
- 00:35:13that it's along the heart somewhere
- 00:35:14along the horizontal axis but it's
- 00:35:16usually randomly placed it's either on
- 00:35:18the right or on the left depending on
- 00:35:19what term you're using but it's clearly
- 00:35:21not correct so the hypothetical here is
- 00:35:24that if you know your left and right if
- 00:35:26you have a long term representation of
- 00:35:28these things and you just bring it to
- 00:35:29bear in in the task then
- 00:35:32kids who do better on left-right should
- 00:35:34also do better in the matching task that
- 00:35:36turns out not to be true this is the
- 00:35:39matching task accuracy over left and
- 00:35:41right production accuracy so what you
- 00:35:43can see is that the four-year-olds some
- 00:35:44of whom are getting 0% correct and some
- 00:35:47of whom are getting 100% correct are all
- 00:35:50matching at about 80 percent so it
- 00:35:52doesn't matter whether you have the
- 00:35:54ability to produce these terms
- 00:35:55accurately in the post-test it also
- 00:35:58doesn't matter if you understand them in
- 00:36:00the post tasks so again we've got this
- 00:36:02no correlation at all between the extent
- 00:36:05to which you know your left's and rights
- 00:36:06and your performance in the matching
- 00:36:08tasks okay so that what this means to us
- 00:36:12is that they don't actually have a
- 00:36:15long-term representation of these terms
- 00:36:16that are they're bringing to bear but
- 00:36:18they're able to use it right in the
- 00:36:20moment when we say to them the red is to
- 00:36:22the left of green they're probably
- 00:36:24coding it somehow as so that's that's
- 00:36:27the one the red is the one that's called
- 00:36:29left at that moment but I'll say more
- 00:36:31about what the format must be so now
- 00:36:34what's the nature of the linguistic
- 00:36:36representation suppose syntactic and
- 00:36:38lexical and so red is left over Ian
- 00:36:41works but red is touching green does not
- 00:36:43work so to explain this difference I'm
- 00:36:46going to argue you need both the
- 00:36:47syntactic frame X is left of Y and you
- 00:36:51need the value the lexical item the
- 00:36:53directional value of left so let's start
- 00:36:56with the syntactic frame this specifies
- 00:36:59figure-ground roles and thereby alters
- 00:37:02the prominence of the two entities it
- 00:37:04tells you which is the figure in which
- 00:37:06is the reference object so if I say red
- 00:37:08is left of green that means the green is
- 00:37:10the ground object and X and and red is
- 00:37:13the figure in the figure object place
- 00:37:15this is actually a widely noted and and
- 00:37:21understand and fairly well understood
- 00:37:22phenomenon lilah Lightman has done a lot
- 00:37:25of work the linguist Lynne talmy has
- 00:37:27done a lot of work the psychologist Amos
- 00:37:31Tversky actually used this in some early
- 00:37:33work on similarity and eleanor rosch has
- 00:37:35has also done work on this and here are
- 00:37:38some examples that we can give so
- 00:37:40there's a difference between saying
- 00:37:42North Korea is similar to China
- 00:37:44and China is similar to North Korea one
- 00:37:46is preferred to the other one sounds
- 00:37:48more natural to the other and that's
- 00:37:50because North Korea is similar to China
- 00:37:52sounds more natural because China is the
- 00:37:54natural at least at this moment China is
- 00:37:56the natural ground or reference object
- 00:37:58in North Korea is compared to it in some
- 00:38:01respects the example from Lin tell me is
- 00:38:05Clark Kent is Superman is better than
- 00:38:08Superman is Clark Kent why because
- 00:38:11Superman is the thing being compared to
- 00:38:13and Clark Kent is the figure object
- 00:38:16being compared to Superman and then
- 00:38:19finally the glytamins did this great
- 00:38:20experiment where they gave people
- 00:38:22nonsense sentences like the zoom met the
- 00:38:25DAX
- 00:38:26if they meets B then B meets a as well
- 00:38:28but they asked people later to rate the
- 00:38:31properties of the zoom and the DAX
- 00:38:32nonsense things and they found out that
- 00:38:34the DAX was old evaluated as older
- 00:38:37bigger more famous and more important
- 00:38:39okay so the effects of the syntactic
- 00:38:42asymmetry what's in figure and what's in
- 00:38:44grand show ZUP in commonplace usage and
- 00:38:46I hope there are fans of Saturday Night
- 00:38:47Live here because we'll have a few
- 00:38:49pieces of data from them the first
- 00:38:52example has to do with Alec Baldwin and
- 00:38:55our president and so as you know Alec
- 00:38:59Baldwin has been playing the role of the
- 00:39:01president on Saturday Night Live and
- 00:39:03this gave me the opportunity to do a
- 00:39:05google search on Alec Baldwin looks like
- 00:39:08Donald Trump versus Donald Trump looks
- 00:39:11like Alec Baldwin okay so in principle
- 00:39:14if Donald Trump is the comparator that
- 00:39:17Alec Baldwin is being compared to you
- 00:39:19should have a larger number of hits it's
- 00:39:21just like the zoom method acts with the
- 00:39:24tax being the the older and more
- 00:39:26important so indeed there are more
- 00:39:28instances of Alec Baldwin looks like
- 00:39:30Donald Trump 700,000 compared to 497 of
- 00:39:34the other variety because this is
- 00:39:36psychological the Association for
- 00:39:38psychological science of course it's a
- 00:39:40replication so it was a replication of
- 00:39:42an earlier finding it also showed up in
- 00:39:452016 so I had done a search at the time
- 00:39:47on Larry David looks like Bernie Sanders
- 00:39:50and it turns out the hit is 22 million
- 00:39:52whereas Bernie Sanders looks like Larry
- 00:39:54David is only 5 million so again you get
- 00:39:56this here
- 00:39:57a symmetry in the way that the thing is
- 00:39:59framed and finally the original finding
- 00:40:01is with of course Tina Fey and Sarah
- 00:40:03Palin Tina Fey looks like Sarah Palin
- 00:40:05two million hits Sarah Palin looks like
- 00:40:08Tina Fey only two hundred and eighty
- 00:40:10eight thousand okay so it's a big
- 00:40:12phenomenon it's also true in addition to
- 00:40:15the frame that the lexical content
- 00:40:16matters so this sets up using left sets
- 00:40:20of a very strong non-reversible property
- 00:40:23so if the X is left of Y then Y is not
- 00:40:27left of X and so so it's impossible to
- 00:40:29if X is left of Y then Y can't be left
- 00:40:33of X it has to be not left of X so it's
- 00:40:35kiley asymmetrical so combining the
- 00:40:38frame idea with the lexical content we
- 00:40:41have come to a very counterintuitive
- 00:40:42prediction which was actually it was a
- 00:40:44great challenge because a linguist in
- 00:40:47our department suggested that this would
- 00:40:48be a prediction if we really believed it
- 00:40:50was frames plus the lexical value of an
- 00:40:52asymmetric term if this non
- 00:40:55reversibility the directionality of that
- 00:40:57item matters then if we say to the kids
- 00:40:59red is prettier than green that should
- 00:41:02work too so we bit the bullet and we did
- 00:41:05the experiment in an experiment five we
- 00:41:07had three conditions one was a no label
- 00:41:09condition the second was the directional
- 00:41:11the Reds on the left right top bottom
- 00:41:13and the third was the red is prettier
- 00:41:15nicer lighter than the green which one
- 00:41:18is exactly the same as the one you just
- 00:41:19saw three different groups of subjects
- 00:41:21and here's what happens no label is
- 00:41:24still in the 60s directional is again in
- 00:41:26the around eighty and prettier is right
- 00:41:29up there with directional left and right
- 00:41:31and so this asymmetrical concept
- 00:41:34prettier actually works for
- 00:41:36four-year-olds to carry over the binding
- 00:41:38of the color location in this very very
- 00:41:40local context we also did difference
- 00:41:43that everything has been a one-second
- 00:41:45delay we did a four second delay it
- 00:41:47turns out the directional terms the
- 00:41:48left's and rights really have a
- 00:41:49persisting value whereas the prettier
- 00:41:52terms decline a little bit but I think
- 00:41:55that's just a detail I want to say
- 00:41:57importantly there's also a developmental
- 00:41:59story to this this line of experiments
- 00:42:02we did the work with three-year-olds
- 00:42:04they can't take advantage of left or
- 00:42:06prettier so we did the no label
- 00:42:08condition the directional condition
- 00:42:10prettier condition these are not
- 00:42:11reliably different from each other
- 00:42:13they're a little bit lower than the than
- 00:42:15the four-year-olds they're at fifty
- 00:42:17percent to fifty five instead of sixty
- 00:42:18percent but there's no effect of the
- 00:42:20linguistic context six-year-olds need
- 00:42:23not take advantage of left and prettier
- 00:42:26actually interferes so six-year-olds
- 00:42:28actually are getting way up high whether
- 00:42:30they have a label or not or or given a
- 00:42:33directional information and so we think
- 00:42:35that what they're doing is basically
- 00:42:36automatically labeling not consciously
- 00:42:38we're not talking about a conscious
- 00:42:40process here but it's an automatic
- 00:42:41labeling and adults can't help but take
- 00:42:44advantage so we did the obligatory task
- 00:42:49the same task with a secondary shadowing
- 00:42:51task in which you have to be talking and
- 00:42:53making horrible decisions along as
- 00:42:55you're doing the task it's actually
- 00:42:56incredibly hard to do this if you're a
- 00:42:58participant so we have no shadow
- 00:43:00condition and then there are verbal
- 00:43:02shadow conditions where you have to
- 00:43:03repeat after you're hearing these terms
- 00:43:06in your earpiece and you'll have to
- 00:43:07repeat as you hear them left right top
- 00:43:09bottom north east south west and so
- 00:43:11forth so you're giving yourself spatial
- 00:43:13terms then we had non spatial terms so
- 00:43:16note meant finished small good you're
- 00:43:19hearing these in your earbud and you
- 00:43:20have to repeat them as you're doing the
- 00:43:21task and then a rhythm shadow where
- 00:43:23you're just shadowing or tapping the
- 00:43:26rhythm that you hear and here's what
- 00:43:28happens the no shadow condition is way
- 00:43:31up at ceiling that's because I think the
- 00:43:33adults are just doing it obligatory and
- 00:43:35automatically rhythm shadow is also way
- 00:43:38up there which means that the verbal
- 00:43:39it's not a verbal interference but the
- 00:43:41verbal shadowing conditions both do
- 00:43:43interfere with the with the performance
- 00:43:47okay so what does language do I'm gonna
- 00:43:51say it's a momentary change it's right
- 00:43:52in the context of the experiment because
- 00:43:54those kids just don't know their left's
- 00:43:56and rights at the end after the task the
- 00:44:00linguistic information is syntactic the
- 00:44:02frame matters its lexical the term
- 00:44:04matters and it's very very abstract
- 00:44:06because prettier also works and how
- 00:44:09powerful is it it's very powerful is
- 00:44:11what I'm going to say and I'm going to
- 00:44:13have a read a quote from George Miller
- 00:44:15in 1956 when he was talking about verbal
- 00:44:19recoding in a different context in the
- 00:44:21context that had to do with
- 00:44:22memory functions and chunking items into
- 00:44:25units but he said if you think of this
- 00:44:27verbal recoding merely as a mnemonic
- 00:44:30trick for extending the memory span you
- 00:44:33will miss the more important point
- 00:44:34that's implicit in nearly all such a
- 00:44:36mnemonic devices the point is that
- 00:44:38recoding is an extremely powerful weapon
- 00:44:41for increasing the amount of information
- 00:44:43in this case I think it's the type of
- 00:44:45information that we can deal with in one
- 00:44:47form or another where recoding
- 00:44:49constantly in our daily behavior so does
- 00:44:52language change thought and if so how
- 00:44:55well I said there are two two views ones
- 00:44:59definitely ones not definitely
- 00:45:00definitely not my view is in the middle
- 00:45:03I'm going to say that language can
- 00:45:05recode it does very frequently by this
- 00:45:07powerful formalism it's not just a
- 00:45:10single word its syntax as well it'd be
- 00:45:12effects are momentary it becomes
- 00:45:15automatic more automatic throughout
- 00:45:18development and it confers an advantage
- 00:45:20for cognition more generally so the
- 00:45:23point I want to make here is that this
- 00:45:25is sort of following DeHaan and cones
- 00:45:27neural recycling hypothesis which is
- 00:45:30that you start out with one kind of a
- 00:45:31visual system a visual system that
- 00:45:33actually doesn't care about reflections
- 00:45:34so this is the same Cup whether it's
- 00:45:37facing right words or left words we
- 00:45:39don't care about the directional facing
- 00:45:41for a visual image of an object but we
- 00:45:44do have to come to the point where we do
- 00:45:46care about it when we do for example the
- 00:45:48learning of letters so we have to be
- 00:45:50able to distinguish peas from B's from
- 00:45:51DS so although the object system isn't
- 00:45:54prepared to encode these differences
- 00:45:57among the reflections once you get
- 00:45:59language in there it can mark those
- 00:46:01representations and distinguish them
- 00:46:03however importantly this does not change
- 00:46:07the non linguistic representation so I
- 00:46:08want to remind you that US adults who
- 00:46:11have full fledged language and use it
- 00:46:12all the time and so forth when I gave
- 00:46:14you the search tasks you acted like
- 00:46:15everybody else acts if it's a single
- 00:46:18feature it's a pop-out if it's a
- 00:46:20combined feature it's hard and it takes
- 00:46:22time so that's where language comes in
- 00:46:24and can amplify I'm gonna just note that
- 00:46:27momentary effects of language are
- 00:46:29ubiquitous I'm actually not going to go
- 00:46:30through these because I think due to
- 00:46:32time but there are many many cases in
- 00:46:35field where things that are that are
- 00:46:37reported as massive effects of language
- 00:46:39on thought actually are momentary
- 00:46:42effects this goes for the color
- 00:46:43literature the space literature and many
- 00:46:45others and I'm going to return to the
- 00:46:47original hypotheses version one two so
- 00:46:50the classic war fee and hypothesis does
- 00:46:53language reshape non linguistic thought
- 00:46:55my answer is no it does not all effects
- 00:46:59are momentary it confers a huge
- 00:47:01advantage but the effects are momentary
- 00:47:03the non linguistic categories are just
- 00:47:05left intact this is true for color
- 00:47:07history for the domain of space and it's
- 00:47:09true for a category of linguistic
- 00:47:11markings called evidential z' the degree
- 00:47:13the degree to which you know the source
- 00:47:15of your information and i want to just
- 00:47:17comment for those of you who are
- 00:47:18interested in this literature you should
- 00:47:20go back and read k and Kempton from 1984
- 00:47:22because they actually say exactly the
- 00:47:25same thing not in these terms but they
- 00:47:26do not claim that there are changes in
- 00:47:28perceptual discrimination quite the
- 00:47:30contrary they show that it's not a
- 00:47:32change in perceptual discrimination and
- 00:47:34I think this has gotten lost in the mix
- 00:47:37so version 2 radical transformation does
- 00:47:40language radically transform thought I
- 00:47:42think in a way yes it records the via
- 00:47:46this formalism that's advantageous to
- 00:47:48cognition it allows you to mark things
- 00:47:50that you wouldn't be able to in your non
- 00:47:52linguistic system therefore having a
- 00:47:54language does radically transform
- 00:47:56cognition and the syntactic frame is one
- 00:47:58of the examples that helps us to change
- 00:48:01and create new representations there are
- 00:48:04many other ways in which syntactic
- 00:48:05frames have their effect I'll give you
- 00:48:07some examples so in source goal
- 00:48:09assignment if something is moving from A
- 00:48:11to B it's both true that they moved from
- 00:48:13A to B and visually you're not able to
- 00:48:16distinguish between what it is that
- 00:48:18you're focusing on but language provides
- 00:48:20you a way of focusing on whether you're
- 00:48:22focusing on the goal or the source
- 00:48:24learning the meanings of novel verbs the
- 00:48:27difference between Jane blick the baby a
- 00:48:29novel a novel verb and Jane blick even
- 00:48:32toddlers can actually use the syntactic
- 00:48:35frame to decide whether or not it's a
- 00:48:37causal event or a non-causal event Jane
- 00:48:40doing something to the baby or Jane just
- 00:48:42doing something on her own political
- 00:48:45persuasion the way that we frame these
- 00:48:48things
- 00:48:48North Korea is similar to China versus
- 00:48:50China is similar to North Korea huge
- 00:48:52differences in the framing now as Lila
- 00:48:55Gladman has pointed out many times this
- 00:48:57is not about the the idea of similarity
- 00:48:59which is what Tversky
- 00:49:01thought that it was it's about the
- 00:49:04language and the information that is
- 00:49:06carried by the language frame these are
- 00:49:08different ideas legal responsibility the
- 00:49:12bus collided with the bus collided with
- 00:49:14the bike versus the bus and the bike
- 00:49:16collided so if the bus collided with the
- 00:49:19bike it's probably going to be the buses
- 00:49:21company the pay's but if it's the bus
- 00:49:23and the bike collided it's probably
- 00:49:26going to be no-fault so you're able to
- 00:49:28derive different kinds of inferences
- 00:49:30from using these two different frames
- 00:49:31and a recent observation from chestnut
- 00:49:34and Markman that I think is quite
- 00:49:35profound gender biases so how many times
- 00:49:38do we say girls are as good at math as
- 00:49:40boys that's putting boys in the
- 00:49:42reference position and girls in the
- 00:49:45figure the girls are being compared to
- 00:49:46the boys and that immediately say it
- 00:49:49sets up an asymmetry my final word is
- 00:49:52this
- 00:49:53the structural properties of language
- 00:49:55have been shown to be crucial for the
- 00:49:57case of the blind child who learns the
- 00:49:59meanings of look and see the structural
- 00:50:04properties of language are also crucial
- 00:50:05for the case of the sighted child who's
- 00:50:07rich perceptual systems have surprising
- 00:50:10limits with respect to products of the
- 00:50:12mind and I just gave you the example of
- 00:50:14the figure-ground case the red and the
- 00:50:16green case but they're also some really
- 00:50:18interesting cases out there in the
- 00:50:20literature one is on number from spy pen
- 00:50:22goldin-meadow
- 00:50:23carry and spell key and others that
- 00:50:25shows that home signers that grew out on
- 00:50:27home signers who do not have have full
- 00:50:29fledged
- 00:50:29numerical system in their sign system
- 00:50:32actually performed differently on number
- 00:50:35number tasks in which they're shown to
- 00:50:38do more estimation and less precise
- 00:50:40numerical judgments and finally the idea
- 00:50:43of theory of mind there's a group jill
- 00:50:45de villiers and others that have argued
- 00:50:47that learning about other people's minds
- 00:50:50is actually deeply entwined with the way
- 00:50:52that we express it in language and
- 00:50:54especially for those of you who know
- 00:50:56about language our linguist or psycho
- 00:50:58linguist in the in the audience it's a
- 00:51:01the compliment structure that mental
- 00:51:03verbs take compliment structures that
- 00:51:05specify that you're talking about the
- 00:51:07contents of mind rather than anything
- 00:51:08out there and so the same lesson applies
- 00:51:14to the blind child and to the sighted
- 00:51:15child and I just want to give a nod to
- 00:51:17the Future these are ongoing studies in
- 00:51:19my lab I just said that the structure of
- 00:51:22the language is crucial for the case of
- 00:51:24the sighted child who's rich perceptual
- 00:51:26systems have surprising limits with
- 00:51:28respect to products of mine and the
- 00:51:30really fascinating set of studies that
- 00:51:32are now ongoing in the lab have to do
- 00:51:34with sighted children's understanding
- 00:51:37about what the verb si means it's
- 00:51:40remarkably limited relative to what you
- 00:51:43think because all of these children see
- 00:51:45however there are very very different
- 00:51:47judgments by young children about
- 00:51:49whether or not a blind person can see
- 00:51:52when do you use the word see turns out
- 00:51:54that blind adults are perfectly happy
- 00:51:57with using the verb see for their own
- 00:51:59activities but children are much
- 00:52:02narrower in their interpretation and
- 00:52:04finally the syntax plays a huge role so
- 00:52:06we're studying when young children
- 00:52:09understand when people see that
- 00:52:12something so the example I'll give you
- 00:52:14is the dog runs there's somebody's
- 00:52:17baking cookies they go out of the room a
- 00:52:19dog comes in and eats up all the cookies
- 00:52:21the person comes back in and says oh my
- 00:52:23god you're eating the cookies and the
- 00:52:25dog runs away leaving crumbs on the
- 00:52:27floor two different questions did the
- 00:52:30girl see the dog eating the cookies yes
- 00:52:33if she was present did the girl see that
- 00:52:37the dog ate cookies you can actually
- 00:52:39infer that the answer could be yes if
- 00:52:41you only saw promise on the floor but
- 00:52:44this is a developmental accomplishment
- 00:52:46to begin to use see in this much broader
- 00:52:47sense that has to do with examining
- 00:52:51aspects of inference that we use in
- 00:52:53seeing and with that I'll close I want
- 00:52:56to give acknowledgments to everybody in
- 00:52:58my lab this is not everybody in my lab
- 00:52:59throughout the years but many many
- 00:53:01people who worked on this problem and
- 00:53:02many related problems this is my annual
- 00:53:04soccer croquet party in the backyard
- 00:53:07along with now deceased dogs two
- 00:53:10Dalmatians but the the biggest
- 00:53:14acknowledgment what I want to give is to
- 00:53:16Lila Lightman who I started out with and
- 00:53:19I think we share a lot of ideas together
- 00:53:21and certainly the idea of asymmetry is
- 00:53:23one that she holds very very dear and I
- 00:53:25think she'd be happy that I was talking
- 00:53:26about this work thank you
- 00:53:29[Applause]
- Barbara Landau
- language and cognition
- Susan Goldin-Meadow
- recoating hypothesis
- cognitive development
- language influence
- thought processes
- syntactic structures
- blind children
- empirical research