To grok or not to grok
Robert Heinlein made up the word grok for his book Stranger in a Strange Land. A character in the book says, "Grok means to understand so thoroughly that the observer becomes a part of the observed." But can everything be grokked? As I write this blog, I am waiting for a relatively complicated Matlab model to finish running (in fact, I have three separate instances of Matlab running on two different computers). When I am done running my models, I will have generated many gigabytes of data. Can I have any hope of grokking that massive amount of data?
Colin McGinn asks an analogous question about "consciousness" in his 1989 article "Can We Solve the Mind-Body Problem?" He concludes no:
"We have been trying for a long time to solve the mind-body problem. It has stubbornly resisted our best efforts. The mystery persists. I think the time has come to admit candidly that we cannot resolve the mystery. But I also think that this very insolubility -- or the reason for it -- removes the philosophical problem."
He goes on to introduce the idea of cognitive closure. Basically, a mind is cognitively closed with respect to a particular problem if and only if it cannot grasp theproblem (one might say cannot grok the problem). He argues that human minds are cognitively closed on the problem of our own consciousness. After all, how could a brain go through any set of cognitive processes to understand itself? "[A]s traditional theologians found themselves conceding cognitive closure with respect to certain of the properties of God, so we should look seriously at the idea that the mind-body problem brings us bang up against the limits of our capacity to understand the world."
Now, I have major reservations with McGinn's approach and conclusions. For example, sometimes questions cannot be answered because they are formulated incorrectly or are fundamentally nonsensical rather than because they cannot be answered within a certain structure (vis Godel's incompleteness theorem). I suspect consciousness is just one of those muddy concepts about which people make nonsensical claims without realizing they are doing so. Also, mathematicians frequently prove difficult theorems only because they can write down intermediate results; should we reject such mathematical proofs because they do not spring full-formed from the mind of the mathematician, as Athena from the head of Zeus? Nevertheless, I find McGinn's idea intriguing and it has shed some light on my approach to some vexing problems.
In a very real sense, I am letting the computer do some thinking for me as I run and interpret the model I am currently working on. This is especially true if (as is true of my model) there is a random, stochastic character to the model. I can make statistical statements about the output of my model, but I cannot hope to atomistic-ally understand my model's output. Does it really matter that I did not personally create the computer's output, so long as I understand how I am using it? Consider a mathematician who uses another mathematician's theorem (or even his own from some time in the past) which he does not currently grok -- is the situation really so different from mine, where the computer is creating the intermediate result? What if the software package being used is not Matlab, but Mathematica, which is explicitly designed to do symbolic calculations? This is a vexing question.
There is another problem of similar ilk that vexes me: genetic "mapping." In his book The Journey of Man, Spencer Wells describes how "tags" on the human genome let us map the dispersal of our species across the globe from our ancestral homeland in Africa. The data Spencer Wells and his colleagues use currently stretches the ability of the scientific community to "grok" and present in a coherent story. I suspect that in the very near future, the data stream will be so massive that only computers will be able to interpret it (if that is not already so). After all, the human genome contains about three billion base pairs and there are currently six billion or so people on the earth! Furthermore, most questions require some sort of relational information (i.e., networks of related people), so that the important questions are really of the order of the exponent of the product of three billion and six billion!!! (FYI, that's a very large number.)
In conclusion, I feel strongly that one of the most important questions facing today's scientists is how to "update" the "scientific method" to solve problems intractable to the individual human mind, or even to the collective scientific community. BTW, I put "scientific method" in quotes because, like consciousness, it is one of those concepts people frequently use in a muddy way. This is unfortunate and (probably) unnecessary. I shan't take the issue up now, but perhaps sometime soon I will post about Quine, Popper (shudder), semantic holism, quantum measurement, and so forth!
Colin McGinn asks an analogous question about "consciousness" in his 1989 article "Can We Solve the Mind-Body Problem?" He concludes no:
"We have been trying for a long time to solve the mind-body problem. It has stubbornly resisted our best efforts. The mystery persists. I think the time has come to admit candidly that we cannot resolve the mystery. But I also think that this very insolubility -- or the reason for it -- removes the philosophical problem."
He goes on to introduce the idea of cognitive closure. Basically, a mind is cognitively closed with respect to a particular problem if and only if it cannot grasp theproblem (one might say cannot grok the problem). He argues that human minds are cognitively closed on the problem of our own consciousness. After all, how could a brain go through any set of cognitive processes to understand itself? "[A]s traditional theologians found themselves conceding cognitive closure with respect to certain of the properties of God, so we should look seriously at the idea that the mind-body problem brings us bang up against the limits of our capacity to understand the world."
Now, I have major reservations with McGinn's approach and conclusions. For example, sometimes questions cannot be answered because they are formulated incorrectly or are fundamentally nonsensical rather than because they cannot be answered within a certain structure (vis Godel's incompleteness theorem). I suspect consciousness is just one of those muddy concepts about which people make nonsensical claims without realizing they are doing so. Also, mathematicians frequently prove difficult theorems only because they can write down intermediate results; should we reject such mathematical proofs because they do not spring full-formed from the mind of the mathematician, as Athena from the head of Zeus? Nevertheless, I find McGinn's idea intriguing and it has shed some light on my approach to some vexing problems.
In a very real sense, I am letting the computer do some thinking for me as I run and interpret the model I am currently working on. This is especially true if (as is true of my model) there is a random, stochastic character to the model. I can make statistical statements about the output of my model, but I cannot hope to atomistic-ally understand my model's output. Does it really matter that I did not personally create the computer's output, so long as I understand how I am using it? Consider a mathematician who uses another mathematician's theorem (or even his own from some time in the past) which he does not currently grok -- is the situation really so different from mine, where the computer is creating the intermediate result? What if the software package being used is not Matlab, but Mathematica, which is explicitly designed to do symbolic calculations? This is a vexing question.
There is another problem of similar ilk that vexes me: genetic "mapping." In his book The Journey of Man, Spencer Wells describes how "tags" on the human genome let us map the dispersal of our species across the globe from our ancestral homeland in Africa. The data Spencer Wells and his colleagues use currently stretches the ability of the scientific community to "grok" and present in a coherent story. I suspect that in the very near future, the data stream will be so massive that only computers will be able to interpret it (if that is not already so). After all, the human genome contains about three billion base pairs and there are currently six billion or so people on the earth! Furthermore, most questions require some sort of relational information (i.e., networks of related people), so that the important questions are really of the order of the exponent of the product of three billion and six billion!!! (FYI, that's a very large number.)
In conclusion, I feel strongly that one of the most important questions facing today's scientists is how to "update" the "scientific method" to solve problems intractable to the individual human mind, or even to the collective scientific community. BTW, I put "scientific method" in quotes because, like consciousness, it is one of those concepts people frequently use in a muddy way. This is unfortunate and (probably) unnecessary. I shan't take the issue up now, but perhaps sometime soon I will post about Quine, Popper (shudder), semantic holism, quantum measurement, and so forth!
posted by MHP at
1:04 AM
0 Comments
