“Some of our categories must originate from another source than direct sensorimotor experience, and here we return to language and the Whorf Hypothesis: Can categories, and their accompanying CP, be acquired through language alone?”
The thing is as we discussed with symbol grounding problem, language can describe and approximate almost anything and we can even keep on refining that description with more words. The only problem is that at some points those words must stop referring from one symbol to the other and actually refer to something that is grounded. This is the symbol grounding problem. So I dont think that categories can be acquired through language alone or else it would mean that hearsay would enable formation of categories. In order to language to create that category it would first need to have a certain amount of words that are grounded, this is done by pointing at something saying this is green for example, in order to base your description on something else than a cycle of symbols.
“Many categories, however, particularly concrete sensorimotor categories (things we can see and touch), are a mixture of the two: categorical at an everyday level of magnification, but continuous at a more microscopic level. Color categories are good examples: Central reds are clearly reds, and not shades of yellow. But in the orange region of the spectral continuum, red/yellow is a matter of degree” It seems like our use of every-day adjectives is an incredibly telling sign of the effect categorical perception has on our description of the world. Many adjectives are indeed, spectrums of phenomena that only make sense to us as one concrete account because of past experiences with/descriptions of that type of thing. It follows that if I am told to “imagine a green field,” there is a whole number of shades, or from a mathematical standpoint, an infinite number of shades I could apply to that field in my minds eye. But that no matter which specific green I couple with the mental image, they will all have the same effect on what it feels like to imagine the field. Of course, it is more than likely that the green that appears to me is some prototypical green, or fluctuating green, based on all my prior experience with green fields. Regardless, I understand a point within an interval of the electro-magnetic spectrum in the same way as I understand any other in this interval, or at least in a way that is similar enough to be irrelevant to my output. This elucidates a dynamic tie between sensory-motor experience and understanding. From here we have indirectly addressed the symbol-grounding problem.
On compression and separation demarcating categorical perception: “Both innate and learned CP are sensorimotor effects: The compression/separation biases are sensorimotor biases, and presumably had sensorimotor origins, whether during the sensorimotor life-history of the organism, in the case of learned CP, or the sensorimotor life-history of the species, in the case of innate CP. The neural net I/O models are also compatible with this fact: Their I/O biases derive from their I/O history. But when we look at our repertoire of categories in a dictionary, it is highly unlikely that many of them had a direct sensorimotor history during our lifetimes, and even less likely in our ancestors' lifetimes. How many of us have seen a unicorn in real life? We have seen pictures of them, but what had those who first drew those pictures seen? And what about categories I cannot draw or see (or taste or touch): What about the most abstract categories, such as goodness and truth?"—I find the case for totally "ungrounded" categories without a direct sensorimotor history, such as the unicorn, to nonetheless be similar to the case of "bachelor", where a more abstract and higher-level category takes on the biases induced by a Boolean combination of categories which are very much grounded (in this case: "horse" and "horn") and gain its own such effects. For example, any horse with so much as a hard, horn-like protrusion would likely be classified as a unicorn—because horses don't have horns! So in this specific case, would it be that fewer unique and normally non-converging tangible categories create sufficient compression and separation that we have no trouble "grounding" this previously ungrounded category?
In this paper, Harnad introduces the concept of Categorical Perception. The important thing about Categorical Perception is it is a categorization on a continuous spectrum.
A continuous spectrum can be, for example, the spectrum of colors. We instinctively divide the spectrum of colors into 7 categories. Of course, there are subcategories for shades of colors. But the most easily perceived categories are only 7. In fact, color is one of the innate categorical perceptions that we possess. The other one is speech perception. But categorical perception can also be learned.
Learning Categorization Perception is also modeled in Artificial Neural Network. Network learns by getting feedback from the environment (e.g. an algorithm or a human supervisor).
Categorization Perception is important to our survival. If we are not able to perceive categories on continuum, we would have a difficult time to do things that we take for granted. For example, colorblind people are not able to perceive the difference between red and green. They would have trouble in front of the traffic light.
"Language-induced CP. Both innate and learned CP are sensorimotor effects: The compression/separation biases are sensorimotor biases, and presumably had sensorimotor origins, whether during the sensorimotor life-history of the organism, in the case of learned CP, or the sensorimotor life-history of the species, in the case of innate CP. The neural net I/O models are also compatible with this fact: Their I/O biases derive from their I/O history. But when we look at our repertoire of categories in a dictionary, it is highly unlikely that many of them had a direct sensorimotor history during our lifetimes, and even less likely in our ancestors' lifetimes. How many of us have seen a unicorn in real life? We have seen pictures of them, but what had those who first drew those pictures seen? And what about categories I cannot draw or see (or taste or touch): What about the most abstract categories, such as goodness and truth?"
In elucidating the distinction between continuous perception vs categorical perception, the paper lists a few approaches that have been taken in order to better understand the role of categorical perception in human categorization. The compresion/separation model seemed most effective to me in explaining how categorical perception occurs. What is most interesting for me is the point captured in the quote above. While it is evident that most categorial perception is rooted in sensory-motor interactions with the world, language appears to have the ability to create categories that are not so easily grounded in that way. The paper points to the possibility the Whorf Hypothesis in attempting to understand how language can do this, however it is not very clear how one might go about understanding language in this way.
In "Categorical Perception" Harnad begins by describing two types of categories: all-or-none and continuous. Some categories fall strictly within the limits of one or the other, however concrete sensorimotor categories have properties of both, being all-or-none at the macro level and continuous when we look closer.
Some categories are innate, meaning we're born knowing them. One example is face detection. In language, speech perception has been found to be based in a continuum of voice onset time. However the way we produce speech naturally seems to be very categorical. Harnad explains this by saying " CP occurs whenever perceived within-category differences are compressed and/or between-category differences are separated, relative to some baseline of comparison."
An important point raised in this paper is that language-induced CP-effects (Whorfian effects) haven't been shown to happen. Some categories are built on others, but those are grounded in sensorimotor experiences and not inferred from language.
Categorical perception and Whorfian hypothesis stand against one another. For Whorf, words were the necessary condition for perceiving differences along a continuum of semantic space. Having numerous words for snow leads to seeing snow differently.
Categorical perception posits that perceiving differences and abstracting invariant features of a set leads to labeling. Snow in different wheather has different features and once you notice that, you label different kinds of snow differently.
To model how children learn words so quickly, Roger Brown came up with a thought experiment called "the original word game." Once a child figures our word boundaries, he has the problem of mapping these arbitrary symbols to their referents. With every new word, the child starts making hypotheses about what the word might refer to. The child procedes to test this hypothesis by using the word to pick out the referents in the set of her working hypothesis. A child here's the word, "potato," and concludes that this new word does not refer to objects for which she already has a word, so it must be something new. She picks up the onion and says, "potato," and is corrected. With more and more feedback, she comes to recognize the difference between a potato and an onion.
In a world where things are not so readily differentiated, words direct are attention not to the differences, but to the fact that there are differences. It is our own task to sort out the invariant features of one category which separate it from another.
not the case that "it is a full-blown language effect, and not merely a vocabulary effect,"
Sometimes new ideas require new vocabulary words. And providing a new word leads us to pick out the invariant features which define that category. But for an entire language to change the way you think, you would need to show that a word in one language lacks features which are provided by its translation in another language.
“Some of our categories must originate from another source than direct sensorimotor experience, and here we return to language and the Whorf Hypothesis: Can categories, and their accompanying CP, be acquired through language alone?”
ReplyDeleteThe thing is as we discussed with symbol grounding problem, language can describe and approximate almost anything and we can even keep on refining that description with more words. The only problem is that at some points those words must stop referring from one symbol to the other and actually refer to something that is grounded. This is the symbol grounding problem. So I dont think that categories can be acquired through language alone or else it would mean that hearsay would enable formation of categories. In order to language to create that category it would first need to have a certain amount of words that are grounded, this is done by pointing at something saying this is green for example, in order to base your description on something else than a cycle of symbols.
“Many categories, however, particularly concrete sensorimotor categories (things we can see and touch), are a mixture of the two: categorical at an everyday level of magnification, but continuous at a more microscopic level. Color categories are good examples: Central reds are clearly reds, and not shades of yellow. But in the orange region of the spectral continuum, red/yellow is a matter of degree”
ReplyDeleteIt seems like our use of every-day adjectives is an incredibly telling sign of the effect categorical perception has on our description of the world. Many adjectives are indeed, spectrums of phenomena that only make sense to us as one concrete account because of past experiences with/descriptions of that type of thing. It follows that if I am told to “imagine a green field,” there is a whole number of shades, or from a mathematical standpoint, an infinite number of shades I could apply to that field in my minds eye. But that no matter which specific green I couple with the mental image, they will all have the same effect on what it feels like to imagine the field. Of course, it is more than likely that the green that appears to me is some prototypical green, or fluctuating green, based on all my prior experience with green fields. Regardless, I understand a point within an interval of the electro-magnetic spectrum in the same way as I understand any other in this interval, or at least in a way that is similar enough to be irrelevant to my output. This elucidates a dynamic tie between sensory-motor experience and understanding. From here we have indirectly addressed the symbol-grounding problem.
On compression and separation demarcating categorical perception: “Both innate and learned CP are sensorimotor effects: The compression/separation biases are sensorimotor biases, and presumably had sensorimotor origins, whether during the sensorimotor life-history of the organism, in the case of learned CP, or the sensorimotor life-history of the species, in the case of innate CP. The neural net I/O models are also compatible with this fact: Their I/O biases derive from their I/O history. But when we look at our repertoire of categories in a dictionary, it is highly unlikely that many of them had a direct sensorimotor history during our lifetimes, and even less likely in our ancestors' lifetimes. How many of us have seen a unicorn in real life? We have seen pictures of them, but what had those who first drew those pictures seen? And what about categories I cannot draw or see (or taste or touch): What about the most abstract categories, such as goodness and truth?"—I find the case for totally "ungrounded" categories without a direct sensorimotor history, such as the unicorn, to nonetheless be similar to the case of "bachelor", where a more abstract and higher-level category takes on the biases induced by a Boolean combination of categories which are very much grounded (in this case: "horse" and "horn") and gain its own such effects. For example, any horse with so much as a hard, horn-like protrusion would likely be classified as a unicorn—because horses don't have horns! So in this specific case, would it be that fewer unique and normally non-converging tangible categories create sufficient compression and separation that we have no trouble "grounding" this previously ungrounded category?
ReplyDeleteIn this paper, Harnad introduces the concept of Categorical Perception. The important thing about Categorical Perception is it is a categorization on a continuous spectrum.
ReplyDeleteA continuous spectrum can be, for example, the spectrum of colors. We instinctively divide the spectrum of colors into 7 categories. Of course, there are subcategories for shades of colors. But the most easily perceived categories are only 7. In fact, color is one of the innate categorical perceptions that we possess. The other one is speech perception. But categorical perception can also be learned.
Learning Categorization Perception is also modeled in Artificial Neural Network. Network learns by getting feedback from the environment (e.g. an algorithm or a human supervisor).
Categorization Perception is important to our survival. If we are not able to perceive categories on continuum, we would have a difficult time to do things that we take for granted. For example, colorblind people are not able to perceive the difference between red and green. They would have trouble in front of the traffic light.
"Language-induced CP. Both innate and learned CP are sensorimotor effects: The compression/separation biases are sensorimotor biases, and presumably had sensorimotor origins, whether during the sensorimotor life-history of the organism, in the case of learned CP, or the sensorimotor life-history of the species, in the case of innate CP. The neural net I/O models are also compatible with this fact: Their I/O biases derive from their I/O history. But when we look at our repertoire of categories in a dictionary, it is highly unlikely that many of them had a direct sensorimotor history during our lifetimes, and even less likely in our ancestors' lifetimes. How many of us have seen a unicorn in real life? We have seen pictures of them, but what had those who first drew those pictures seen? And what about categories I cannot draw or see (or taste or touch): What about the most abstract categories, such as goodness and truth?"
ReplyDeleteIn elucidating the distinction between continuous perception vs categorical perception, the paper lists a few approaches that have been taken in order to better understand the role of categorical perception in human categorization. The compresion/separation model seemed most effective to me in explaining how categorical perception occurs. What is most interesting for me is the point captured in the quote above. While it is evident that most categorial perception is rooted in sensory-motor interactions with the world, language appears to have the ability to create categories that are not so easily grounded in that way. The paper points to the possibility the Whorf Hypothesis in attempting to understand how language can do this, however it is not very clear how one might go about understanding language in this way.
In "Categorical Perception" Harnad begins by describing two types of categories: all-or-none and continuous. Some categories fall strictly within the limits of one or the other, however concrete sensorimotor categories have properties of both, being all-or-none at the macro level and continuous when we look closer.
ReplyDeleteSome categories are innate, meaning we're born knowing them. One example is face detection. In language, speech perception has been found to be based in a continuum of voice onset time. However the way we produce speech naturally seems to be very categorical. Harnad explains this by saying " CP occurs whenever perceived within-category differences are compressed and/or between-category differences are separated, relative to some baseline of comparison."
An important point raised in this paper is that language-induced CP-effects (Whorfian effects) haven't been shown to happen. Some categories are built on others, but those are grounded in sensorimotor experiences and not inferred from language.
ReplyDeleteCategorical perception and Whorfian hypothesis stand against one another. For Whorf, words were the necessary condition for perceiving differences along a continuum of semantic space. Having numerous words for snow leads to seeing snow differently.
Categorical perception posits that perceiving differences and abstracting invariant features of a set leads to labeling. Snow in different wheather has different features and once you notice that, you label different kinds of snow differently.
To model how children learn words so quickly, Roger Brown came up with a thought experiment called "the original word game." Once a child figures our word boundaries, he has the problem of mapping these arbitrary symbols to their referents. With every new word, the child starts making hypotheses about what the word might refer to. The child procedes to test this hypothesis by using the word to pick out the referents in the set of her working hypothesis. A child here's the word, "potato," and concludes that this new word does not refer to objects for which she already has a word, so it must be something new. She picks up the onion and says, "potato," and is corrected. With more and more feedback, she comes to recognize the difference between a potato and an onion.
In a world where things are not so readily differentiated, words direct are attention not to the differences, but to the fact that there are differences. It is our own task to sort out the invariant features of one category which separate it from another.
not the case that "it is a full-blown language effect, and not merely a vocabulary effect,"
DeleteSometimes new ideas require new vocabulary words. And providing a new word leads us to pick out the invariant features which define that category. But for an entire language to change the way you think, you would need to show that a word in one language lacks features which are provided by its translation in another language.
Gr5 titanium stone from a garden | TiG, India | Titanium Arts
ReplyDeleteGr5 titanium stone from a garden, you will use it as a suppliers of metal conduit titanium wheels to titanium pots and pans your inner tree. It revlon titanium max edition can be used micro touch titanium trim where to buy for gardening and production of ceramic wood.
i520w5balqc535 vibrators,dog dildo,dildo,sex chair,dog dildos,anal sex toys,Discreet Vibrators,penis sleeves,vibrators r112t6vkbyr647
ReplyDeleted385p1xxswo207 cheap sex toys,sex toys,Wand Massagers,horse dildo,g-spot dildos,male sex dolls,sex doll,dildos,custom sex doll n147m2vmczm744
ReplyDelete