This is one giant rant about how crazy age limits are in this country.
It began when I was 18. One of my favorite jazz guitarists and arguably one of the best jazz guitarists of all time came to my home town, minutes away from my house. Well, this venue only allowed people 21 and over, so I had to pass on it. At that time, I felt it was somewhat fair that I was jipped out of my chance to see a great performer because I was still a little underage.
Later, the person who caused me to discuss my pain with you, Mr. Victor Wooten, came to that same jazz venue when I was 19. Again I was jipped and could not go. My frustration was still there, but again, I can understand it. For some reason being two years younger is a decent justification.
So now, I am 20 years old, and my birthday is in two months. Mr. Wooten is coming to Bloomington, IN on September the 19th to perform has magic on the bass guitar, and I cannot go because I am 20 years old still. I want justification, America. I want to know why, a person like myself, who is far more astute and responsible when it comes to the consumption of alcohol, who has never been drinking and driving and finds those that do absolutely disgusting and, most importantly, is only ONE MONTH away from being the legal age to enter a bar cannot attend an event.
Here is my problem with age limits: The hard arbitrary link between being a certain age and having a certain privilege is never going to be perfect. We live in a country where you can smoke your lungs away and have guaranteed problems after just one week of smoking, you can gamble all of your money away in lottery tickets, you can vote for your leader, and you can drive a two ton automobile and risk killing yourself and others but you cannot drink a beverage in public. You cannot even go into a place or area where alcohol is served. I understand that there should be some limit clearly, because we would not want 10 year olds drinking or smoking, but then there is still the issue of 9 and a 1/2 year olds…why are they exempt?
Perhaps one way of restricting alcohol use is some signed consent by the parents or guardian and the user to be noted on an ID. If the parents believe that their child is competent enough to drink alcohol, or at least sit in a location where there is alcohol served, the minor should be allowed to assume whatever risks that entails. But the issue here may be that the minors would perhaps “ruin the fun” of the adults by being around. I’m sure older adults want to only associate with other older adults. So in this case, the age limit is really a social thing. If that’s the case, then what stops a guy like me from entering a bar, or others who are just that close?
So then why not make restrictions based on certain groupings? I am a college student on the campus, and I do not think that 21 year old would have a problem with a 20.9 year old sitting there enjoying music. Perhaps in a general area, 21 and up is the way to go, but on a college campus, maybe upper classmen only? This might be stretching because I really just want to be in the in-group. Which is probably the reason why this solution would not go so well, because there would be fights over who is in and who is not.
None of this musing will get me what I want in the end anyway. I really need your help! If you understand my plight and support the fact that the standard does not make sense, post a comment!
May 31, 2009 • 10:31 pm 0
Concepts and Categorization: Commentary on Goldstone 2003
The following is an abstraction and discussion on “classical” cognitive science paper “Concepts and Categorization” (PDF: Goldstone & Kersten 1993). To begin, we have known for a while that “people have a natural tendency to make a thing something” (Wittgenstein 53). Whenever we perceive a novel object, we want to put it somewhere in our brain to remember it later. This tendency is the foundation for categorization. Another thing that we do besides categorizing is forming a concept. Categories and concepts are not the same thing. The concept of [dog] (brackets will be used as the naming convention for concepts) is whatever psychological state that signifies thoughts of dogs. In your head there exists an idea of how a dog should be, and that is the concept of a dog. On the other hand, a category <dog> (arrowheads will be used as the naming convention for categories) is all of the entities that are correctly categorized as dogs. The definition is recursive, but is still useful. While the concept is in your head, the category is actually a collection of objects that have been coined as a dog. One question that pops up is somewhat like a “chicken and the egg” issue: Does a category determine what concepts are formed, or do concepts determine which categories are formed? If it is categories, then concept learning is “inductively creating a mental structure that will predict a category.” In other words, the idea of dog in your head is going to be continuously updated in order to successfully determine if an object is a dog. If it is concepts, then “external categories are the end product of applying concepts to things in the real world.” So the idea of a dog was already in your head and categories are formed based on whether it is in the concept or not. A peculiar feature of concepts is that they make things that are very different similar in some ways. [Clothes] can make t-shirts and pants on an equal plane. In contrast, categories still have entities that are distinguishable. Concepts are a type of equivalence class because of their ability to makes things that look superficially different to be the same. Because of this, concepts that are created are not usually changed if the outsides of objects are different. An example study has shown that if participants are told that a new animal has been discovered that looks like a raccoon but internally has all the features a skunk has, people categorize it as a skunk. Regardless of how things look, if they share properties that are equivalent to an idea in one’s head, then these objects will be clumped in the idea of a concept.
There are several uses for concepts. Concepts function as filters in perception.
Because of this, we can create new ideas about novel objects we have seen by perhaps taking two concepts we have been previously exposed to. Interestingly, this composition of concepts can be viewed in two ways. Goldstone provides an excellent example of this: “How does one interpret the term buffalo paper when one first hears it? Is it paper the shape of a buffalo, paper used to wrap buffaloes presented as gifts, an essay on the subject of buffalo, coarse paper, or something used like fly paper to catch buffalo?” Two models have been presented to represent two ways of looking at how we combine two concepts. In one model (Murphy 1998), one finds a variable in the second noun to be filled by the first noun. So, buffalo paper is paper that is shaped like a buffalo. On the other hand, properties from the first concept may be transferred to the second concept (Wisniewski 1997, 1998). This may be difficult with our buffalo paper example, but perhaps this could be paper that feels like a buffalo. When concepts are combined, the concept’s meanings combine to determine the meaning of the conjunction, but there may be other interactions between the two concepts and the “real-world plausibility” of this novel concept that influence the conjunction’s meaning. In this sense, it seems that we would throw out the idea that buffalo paper is “paper used to wrap buffaloes as gifts,” since in our real-world experience we do not encounter buffalo being presented as a wrapped up present with a bow on top.
In addition to allowing for interesting combinations, concepts also allow for inductive predictions as to what should be in a category. So, an instance of a slobbering dog observed may cause us to conclude that dogs slobber, and so any object that slobbers is more likely to be a dog. This inductive hypothesis making works because members of the category “dog” need not share the exact features across all the entities, but do need to resemble each other, like that of a family (Rosch & Mervis 1975). So, “slobbering” is associated with “wet noses” and “wagging tails.”
Concepts also allow for easy communication of a package of information. When one says “Ed is a football player”, [football player] implies large, in-shape, athletic, and rough-around-the-edges, because the of the nature of the sport and the idea that results from [football]. Because of this ability to transmit a lot of information due to one concept, the same can be seen for categories of information. Categories allow for the storage of particular instances more efficiently, so concepts are useful to conserve memory. If an object can be identified as belonging to a pre-established category, then less cognitive processing is necessary (assimilation vs. accommodation, Piaget 1952).
There are four ways that concepts can be represented in our minds. One way is rule-based, like an entry in a dictionary. This way results from a huge uprising in the science of hypothesis testing (Bruner, Goodnow, Austin 1956) – if the object has these psychological states, then it is this concept. Interestingly, rules that are conjunctives (if the object is white and round) are more easily processed than implication rules (if the object is white, then it is round), which are in turn easier to process than bidirectional rules (the object is white if and only if it is round). Along these lines is the “classical view” triumphed by Fodor (1963). In this view, the word bachelor consists of markers like +male, +adult, +physical object, -married. However, this view has been severely challenged. One example is that the word game is harder to find rules for, since it can mean multiple things. Wittgenstein suggests that it is “better to think of members of a category as being related by family resemblance” Another problem with this view is that category membership may be unclear. This is not just with other entities within the category but also even with their own selves! Finally, not all members of a category are equally good. While bachelor could be +male +adult etc…a counter example shows that the marker “adult” would be much more important than “male”, because a 5-year-old male is also technically a bachelor.
A second way that concepts are represented is making prototypes. In this case, concepts are organized based around family resemblance, rather than features that are individually necessary to the concept. The prototype represents the most common attribute values. This sort of harkens back to Euclid and the idea of a “perfect circle” – there exists in our mind the ideal most common example of an entity that we base against this prototype. This way of concept representation is more intuitive than rule-based, classical views because it can handle unclear categorical representations and can lead to graded members within the category. Along this line of thought is the idea that categories are formed based on a “central tendency” to that which an entity is categorized. Four objects with ratings of 7, 8, 4, 5 out of 10 (10 being most similar to the prototype) would have a central tendency of 6, and so would allow for graded entities.
A third method would be with exemplar theory. This assumes that we keep a track of every instance, for example, of a table in order to decide whether an instance belongs in the table concept. This also assumes that a category represented by exemplars are based on decisions of the similarity of objects categorized; therefore, if a certain table becomes more similar to the tables in our minds, we tend to give it a more likely probability that the object is a table. There is a question as to how we represent these exemplars. One thought is that we use a multi-dimensional space (Hintzman 1986). Think of this as a grid of objects that clump together when they are similar and spread apart when they are not. While this seems intuitive for a human to do, we would need to know how to program this into a computer to perhaps generate a working model for this approach. Moreover, there have been studies to suggest that people perform better at recognizing a novel stimulus as being part of a different category, after being trained on previous stimuli that belong in another category. So if a person is presented with a series of dogs and then is shown a cat, they are more likely to say that a cat is not in <dog>. But according to exemplar theory, we would have needed to see a cat in order to put it in [cat]. Despite these issues, predictions in exemplar theory do better than those in prototype models because exemplars can also predict that familiar distortions will be categorized more accurately than novel distortions that are equally far removed from the prototype. Of course, how can someone store every possibility instance of a category remains to be answered. I cannot say I remember every instance of a table in my mind, and this idea does not seem to allow for only some tables to not be represented.
Finally, concepts can be represented as category boundaries. In contrast to prototypes, category concepts represent the limits of a categories’ scope. People are more likely to distinguish two different categories than two members of the same category, partly because of this approach. When people categorize, either they compare prototypes of a category or use boundaries as reference points. When using models for each of these theories on a test for accuracy, both models perform equally well at distinction. Sometimes, the categorization based on boundaries depends on context. Previous repetitions of instances in one category can cause an instance to be placed in another category.
When taking in all of these approaches, they have all ignored what the concepts actually mean. Although a bike is similar to a car in that they are [transportation devices], a bike requires actual human physical endurance, whereas a car involves the knowledge of being able to drive it. These meanings would heavily influence where an instance would go. It seems that how we categorize depends on our theories of the world. For example, if we saw a man fully clothed jumping into a swimming pool, we may categorize this person as drunk, rather than a swimmer (Murphy & Medin 1985). Forming categories can also depend on statistical evidence along the lines of theories we make about the world. Evidence has shown that we use multiple sets of these theories above (like most sciences that have multiple theories, it seems).
Keeping all of this information in mind, it is important to recognize that concepts can be influenced by other factors, like mentioned above. Furthermore, concepts have a bidirectional influence on language. One’s repertoire of concepts may influence the types of word meanings one learns, whereas the language that one speaks may influence the types of concepts that one forms. Children have been found to learn nouns much sooner than verbs, perhaps due to the ambiguous mapping that verbs have and their meanings being irrelevant to the event actually occurring. Therefore, word meanings should be easier to learn if they can be mapped to existing concepts. Moreover, the existence of a label in association with a category influences if a category is learned or not:
In looking to the future, Goldstone notes that concepts are used both to recognize objects and to ground word meanings. Future research, therefore, should try and hash out this dual nature of concepts, in order to better understand them. Furthermore, research in attempts to model the representation, formation, and usage of concepts should be advanced, particularly in neural network modeling and dynamical systems. Finally, knowledge of concept learning should be used in a more practical way, like educational reform. If we can figure out how concepts are stored and represented, this has a huge benefit in how we teach children the subtle differences and similarities between objects, so they may understand the world at a faster rate.
Selected Bibliography from Goldstone & Kersten 2003:
Bruner, J. S., Goodnow, J. J., & Austin, G. A. (1956). A study of thinking. New York: Wiley.
Fodor, J. A. (1983). The Modularity of the mind: An essay on faculty psychology. Cambrdige, MA: MIT Press.
Goldstone, R.L., & Kersten, A. (2003). Concepts and Categorization. In A.F. Healy & R. W. Proctor Comprehensive handbook of psychology, Volume 4: Experimental psychology. 599-621. New Jersey: Wiley.
Hintzman, D. L. (1986). “Schema abstraction” in a multiple-trace memory model. Psychological review. Iss. 93. 411-429.
Murphy, G. L. (1998). Comprehending complex concepts. Cognitive Science. Iss. 12. 89-115.
Murphy, G. L. & Medin, D. L. (1985). The role of theories in conceptual coherence. Psychological Review, Iss. 92. 289-316.
Rosch, E. & Mervis, C. B. (1975). Family resemblances: Studies in the internal structure of categories. Cognitive Psychology. Iss. 7. 573-605.
Piaget, J. (1952). The origins of intelligence in children. New York: International Universities Press.
Wittgenstein, L. (1953). Philisophical Investigations. (G.E.M. Anscombe, Trans.). New York: Macmillan.
Winsniewski, E. J. (1997). When concepts combine. Psychonomic Bulletin and Review. Iss. 4. 167-183.
Winsniewski, E. J. (1998). Property instantiated in conceptual combination. Memory & Cognition. Iss. 26. 1330-1347
Filed under: Neuro/Cogsci , cogntive science, commentary, concept, equivlance class, exemplar, goldstone, perception language, prototype