Semantics Podcast Flashcards
words are referential
they refer to things in the outside world, they are referring to an idea
word are arbitrary
no dependence between a sound of a word and the meaning (usually, onomatopoeia is an exception: buzz)
signed language as a bit less arbitrary?
better able to take advantage of iconic cues
word-referent relationship is symbolic
the word STANDS for the object it’s referring to
how does a kid figure out what makes a dog? how do they recognize new dogs?
children’s representations consist of semantic features (dogs bark, have four legs, has fur, it pants, doesn’t meow, doesn’t purr)
certain features matter more than others: barking more than four legs
prototypical theories
you learn prototypical concepts (the most common thing) first, then extend to “weirder” stuff
subtypes: (1) essential features that define categories or (2) there aren’t necessary and sufficient features, just more and less likely ones
essential features that define categories
redness is defined: does it reflect in a particular wavelength of the visual light spectrum? if not, then it’s not red
harder for things like dogs, they don’t necessarily have features that are necessary and sufficient, just less likely
kids are learning words and concepts at the same time
may have concepts before they learn words to describe these concepts
big central problem of language development
in any given situation, there are a lot of concepts that a word could apply to
formulated as the gavagai problem
Quine’s problem (the “gavagai” problem)
you’re in a foreign country
one of the residents of the country is standing next to you and a rabbit hops across the road and the resident looks at the rabbit, points at it, and says “gavagai!”
do you really know what a gavagai is?
it could be rabbit, or a proper name, or a rabbit ear, or dinner, or danger, or… HOLY CRAP
the point is, is the child word learner is in this situation (pointing or gaze direction) but there are a large number of potential things that that word could be referring to
joint attention
child uses pointing and gaze direction (or the parent direct’s the child’s attention) and they share attention toward a region of physical space that usually has an object in it
by both of them paying attention to the same thing: maybe that’s how the kid learns the word: “look at that! that’s coffee!”
Association learning view
controversy!!!
very non-specific to language
hearing the word predicts the presence of the object (or concept) and seeing the object predicts the presence of the word
cross-situational word learning
Yu&Smith
maybe the way that kids learn the word “ball” is that they keep track across situations, what are the most frequent co-occurrences of words and objects
kids presented with a scene of a ball and a bat (which is which?)
then kids presented with a scene of a ball and a dog and they hear “ball and dog”
if they’re keeping track across situations, then the probability of the word ball given this particular object will be higher than the other words
trained infants to test this with two nonsense shapes with two nonsense names: if they’re keeping track across multiple objects, they should be able to parse out which word is which: how long did they spend looking at the right object? if they don’t care they should look at both objects for equal amounts of time: they look significantly more often at the correct object
