My OperaA Model of a Student Using the Harvard Architecture
Sunday, September 18, 2005 12:45:00 PM
Basing on the prior experience of the uneasy relationship with my own memory, I've concluded that it is quite non-uniform. If I was to draw an analogy between human and computer memory, I'd rather compare the brain with the Harvard rather than Von Neumann architecture. To remind the reader, the former has separate memory for program and data, and the CPU accesses it in different ways; the latter uses common memory for both program and data.
According to this analogy, the brain contains fundamentally different structures for memorizing pure information (data) and procedures (program). In some people the former is better developed, and in some people, like me, the latter. I suppose that the latter is actually more common because there is a lot of methods to memorize pure information by substituting a generating procedure for it. For example, to memorize a telephone number like 1252627, many would find it easier to remember a procedure: write down 1 and then three consecutive integers starting with 25. I think that most people, while trying to memorize this number, will find the above procedure (without verbalizing it) and memorize it instead of the original number. Not every number can be represented with a reasonably short procedure like this, that's why some telephone numbers are considered easy to remember, and some are harder because one has to use the less developed memory for pure information. However, some people don't need the mnemonic tricks, and they can memorize all telephone numbers equally easy. These people have the memory for pure information better developed than the procedure memory.
I would call the content of the pure information memory knowledge, and the content of the procedure memory comprehension, even though this doesn't entirely agree with the casual use of these terms. Of course, the analogy with computer memory is incomplete: in my conception, the second type of memory holds not only procedures but also generalizations and relations (it's possible that there is some finer subdivision of memory types). Knowledge, on the other hand, is uniform, and there is no subdivision by type of information. To understand some information means to promote it from knowledge to comprehension by analyzing it, finding patterns, and probably relating to the previously understood. The ability to comprehend highly depends on the nature of information and on the similar comprehension already acquired. When someone says they have good memory for numbers, or melodies, or poetry, it means that they are relatively good at analyzing particular kinds of information for transforming them into comprehension. On the other hand, highly developed ability to accumulate knowledge rather than comprehension is more rare. Extreme cases of such development (such as when a person can recite the entire book after reading it once) are usually called infallible memory.
Human brain is a flexible system, and it's often that one system or mechanism in it gets replaced with another. In particular, replacement of knowledge with comprehension happens almost constantly. This incurs a certain overhead for both storage (analysis, pattern identification) and retrieval of information. For example, to call the telephone number memorized as the procedure described above, one has to execute the procedure first. The reverse replacement is also possible: someone who is unable or unwilling to analyze some information can memorize it as knowledge. In situations where the original information has to be reproduced precisely, knowledge is preferable to comprehension because it doesn't require any transformation. On the other hand, it's impossible to use knowledge in situations that require comprehension (though in some cases one can transform stored knowledge into comprehension, in other words, understand it when it becomes necessary).
The hypothesis about human memory using the Harvard architecture is valuable for pedagogics. The goal of teaching is to have the students comprehend the material, but often knowledge is substituted for comprehension. This happens for a variety of reasons: a better capacity for memorizing of pure information rather than for understanding, lack of ability or training for the analysis necessary for comprehension, lack of understanding of the previously learned material which is often crucial for understanding of the new. A great part of the substitution of knowledge for comprehension should be attributed to the imperfect system of progress assessment based on standard tasks and even on simple reproduction of the material. In the latter case, knowledge can be used instead of comprehension to reproduce the information; in the former case, it's impossible, but a student who is unwilling or unable to attain adequate comprehension of the material can rather attain inadequate comprehension. This kind of comprehension uses the same approach as the adequate comprehension but makes only shallow analysis resulting in identification of superficial patterns. An example is a memorized procedure for solving a type of standard problems in school physics: s divided by v yields t. Obviously, inadequate comprehension is as deficient with regard to the goals of teaching as pure knowledge. It is clear that improving the progress assessment system alone won't yield better results but will just lead to deterioration of the average figures. It is necessary to improve the methods of presenting the material as well as of training and stimulation of the abilities for information analysis. The latter is crucial because the analysis necessary for understanding happens in the student's head, and it can't be replaced by any visual teaching methods. There is always the work that the students have to do on their own, and if they cannot or don't want to do it, the teacher's work is wasted.
По-русски: Модель учащегося с гарвардской архитектурой
According to this analogy, the brain contains fundamentally different structures for memorizing pure information (data) and procedures (program). In some people the former is better developed, and in some people, like me, the latter. I suppose that the latter is actually more common because there is a lot of methods to memorize pure information by substituting a generating procedure for it. For example, to memorize a telephone number like 1252627, many would find it easier to remember a procedure: write down 1 and then three consecutive integers starting with 25. I think that most people, while trying to memorize this number, will find the above procedure (without verbalizing it) and memorize it instead of the original number. Not every number can be represented with a reasonably short procedure like this, that's why some telephone numbers are considered easy to remember, and some are harder because one has to use the less developed memory for pure information. However, some people don't need the mnemonic tricks, and they can memorize all telephone numbers equally easy. These people have the memory for pure information better developed than the procedure memory.
I would call the content of the pure information memory knowledge, and the content of the procedure memory comprehension, even though this doesn't entirely agree with the casual use of these terms. Of course, the analogy with computer memory is incomplete: in my conception, the second type of memory holds not only procedures but also generalizations and relations (it's possible that there is some finer subdivision of memory types). Knowledge, on the other hand, is uniform, and there is no subdivision by type of information. To understand some information means to promote it from knowledge to comprehension by analyzing it, finding patterns, and probably relating to the previously understood. The ability to comprehend highly depends on the nature of information and on the similar comprehension already acquired. When someone says they have good memory for numbers, or melodies, or poetry, it means that they are relatively good at analyzing particular kinds of information for transforming them into comprehension. On the other hand, highly developed ability to accumulate knowledge rather than comprehension is more rare. Extreme cases of such development (such as when a person can recite the entire book after reading it once) are usually called infallible memory.
Human brain is a flexible system, and it's often that one system or mechanism in it gets replaced with another. In particular, replacement of knowledge with comprehension happens almost constantly. This incurs a certain overhead for both storage (analysis, pattern identification) and retrieval of information. For example, to call the telephone number memorized as the procedure described above, one has to execute the procedure first. The reverse replacement is also possible: someone who is unable or unwilling to analyze some information can memorize it as knowledge. In situations where the original information has to be reproduced precisely, knowledge is preferable to comprehension because it doesn't require any transformation. On the other hand, it's impossible to use knowledge in situations that require comprehension (though in some cases one can transform stored knowledge into comprehension, in other words, understand it when it becomes necessary).
The hypothesis about human memory using the Harvard architecture is valuable for pedagogics. The goal of teaching is to have the students comprehend the material, but often knowledge is substituted for comprehension. This happens for a variety of reasons: a better capacity for memorizing of pure information rather than for understanding, lack of ability or training for the analysis necessary for comprehension, lack of understanding of the previously learned material which is often crucial for understanding of the new. A great part of the substitution of knowledge for comprehension should be attributed to the imperfect system of progress assessment based on standard tasks and even on simple reproduction of the material. In the latter case, knowledge can be used instead of comprehension to reproduce the information; in the former case, it's impossible, but a student who is unwilling or unable to attain adequate comprehension of the material can rather attain inadequate comprehension. This kind of comprehension uses the same approach as the adequate comprehension but makes only shallow analysis resulting in identification of superficial patterns. An example is a memorized procedure for solving a type of standard problems in school physics: s divided by v yields t. Obviously, inadequate comprehension is as deficient with regard to the goals of teaching as pure knowledge. It is clear that improving the progress assessment system alone won't yield better results but will just lead to deterioration of the average figures. It is necessary to improve the methods of presenting the material as well as of training and stimulation of the abilities for information analysis. The latter is crucial because the analysis necessary for understanding happens in the student's head, and it can't be replaced by any visual teaching methods. There is always the work that the students have to do on their own, and if they cannot or don't want to do it, the teacher's work is wasted.
По-русски: Модель учащегося с гарвардской архитектурой
