006a-messaging, reporting and shared simulation
messaging and translation - we see shared experience as the basis for translation.
Can we show an external universe, that is a universe that is external of representations, using messaging? Could we test externality by leaving messages? We think of messages as being external. We create messages with the implicit belief that they are in fact external to us. Messages are things we project, express, or leave for others to encounter and experience.
Do the assumptions in our actions mean that messages are, in fact, external?
Messages, especially made to last over time, are always treated subjectively. The meaning of a message is "internal" and not part of a physical external universe. ie, the medium is not the message.
The material components or medium of a message are external and objective (in the ordinary sense). But the meaning of a message is not. Meaning is subjective. :) is that a smiley face or is it a colon and a closed parenthesis?
In computer parlance, a message is an external object. A message is really nothing more than an input to some function. A typical computer system does not "understand" a message. It processes a message and produces outputs based on the program used to process the input data. It is the programmer who understands the message. At best a computer can receive and possible create programs or algorithms to process messages. But usually it is the programmer that writes programs to process messages. Even computer systems that generate their own programs for message processing are programs that are written by programmers. There are no programs that self-generate both rules and messages.
This is true for neural networks or evolutionary programs. Because, the computer is invariably working with variations in the medium of a message. And not really dealing with the content of the message at all. Human beings work with both the content of messages and the medium of messages. Content, at least today, is a mystery to programs.
Messages are representations using a medium. Computers have been designed to deal with variations in a medium. Say, scanning in text and converting it to words. but the computer does not understand the words. the computer cannot understand a message if say the text went from 12pt to 300pt. But a person could, because they are looking at the words as conveying a message.
for instance, a computer would not understand that :
But if you can read it, you can understand it. And you know that the math is wrong.
Searle's famous chinese box explains the problem with message passing quite well. But message passing is used as an argument for an external, representation free universe. The medium of a message is not what is relevant to the reader, but it is all that is available to a computer. The universe contains not just messages with data, but also content with meaning. The meanings exist in the universe as well.
The ink on paper is irrelevant to an ordinary reader. it's the message that is important. Learning about messages requires learning about mediums, about forms and shapes. Learning to read, and learning letters will provide us insight into this problem. A computer never learns letters. Ascii (or hexadecimal or binary) is not something a computer learns, it is the data of computer processes and messages. Do human beings have a symbol set that underlies cognitive function?
Computers are not designed to work with the associative aspects of letter symbols. With OCR (Optical Character Recognition) A computer does the recognition tasks, but none of the associative tasks. For example: letters are associated with sounds, and shapes, and sequences. Letters can be formed, and spoken and written, and written backwards and wrong. A letter is the same regardless of the font or it's size. A letter is a pure abstraction. Letters in this regard have a number of platonic qualities. It doesn't matter how a letter appears, it is still a letter. **see phenomonalism and platonic forms
But to an ordinary computer, a letter is actually a binary string. It is not a letter at all. Because a computer does not understand letters, a computer cannot read.
When a person claims a computer can read, they do not really mean that at all. they mean that a computer is capable of taking some images and converting those images to binary strings which when presented to a viewer as an image of letters or numbers we can see the computer converted the images to what we see as the correct letters and words. And if the computer "sees"
and converts it to 2 + 2 = 22 we should not think the computer did anything wrong. clearly the software needs to be tweaked to better recognize variations in different symbols. the computer got nothing wrong, it simply followed it's program. And it is only after we look at a computer's output that we can analyze if it's output is wrong. Most computer systems do not read their own outputs.
If a computer "saw" it's own output, how would it ascertain if it as in error? Does a computer make an association of the ASCII characters 2Z back to the image
like you can?
Computers produce outputs but they do not report on their output. A computer does not engage in self-reporting like a person does. A computer does not say, "I couldn't quite tell if this was '22' or '2Z'." If a computer does produce a listing of areas where it might have two options, it may produce what we call a report, But it is the programmer who engaged in the analysis of those kinds of events and wrote the report. Or the program created an abstraction of the report where the computer "fills-in" the areas where there might be multiple possibilities in the character recognition. But the computer is not self-reporting.
When people self report, a person converts the small amount of input information conveyed and associates it to their own experiences and ideas to make sense of reports. The reason that a person MUST do this is because reports are made up of small amounts experiences.
When we read a story or a novel, what we see are ink blotches scattered all over a page. This visual information we associated to letters, and into words, and words into sentences. The words and sentences we associate to sights, sounds, tastes, smells, feelings, ideas. All this association making is subjective. The association making is not what we ordinarily think of as being external. Only the inky blotches, the marks on paper, are external. The visual shapes are what we classify as being external.
Or perhaps it's the shape of negative space that is what is reported in our brains. What do you see?
reports are messages. messages must be interpreted.
reports are minimal data. Consider a list of employees. If we list them with gender values of M and F what does that mean? It means those employees gender contain associations to penises and vaginas. what else do you think M and F mean? When you put your gender on a form, what are you saying? If you are male, when you fill out a form asking for your gender and you choose "M" are you cognitively saying "I have a penis?" Probably not, but it is a fact that is entailed in that message of "M".
reporting and simulation
do changes in weather temperature communicate a message to human beings ( the coming of winter/summer) do they communicate a message to a thermostat? Is temperature, or "mother nature" COMMUNICATING?
No, obviously not. But then why does warming indicate summer and cooling indicate the coming of winter? We are assigning meaning to these experiences. That two people could say "it smells like spring" and be in perfect agreement based on a set of chemicals in the air means the two people share a representation associated to those particular chemical experiences.
An example of poor data passing between people is when you try to tap out a song to a partner. when tapping out a song to another person can the receiving person tell what song the tapper are tapping out?
Usually, the receive cannot tell what song is being tapped out. Because the sender's taps are so strongly associated with the song in the sender's mind that the sender cannot see the data of the tap on it's own terms. The receiving person has a difficult time associating the rhythm of the taps to the melody in the sender's head.
Computers pass tap messages to each other. Computers do no communicate with each other. They pass data to each other. Data is not what is typically passed between humans. Messages, as indications of experiences, are passed that must be imputed or instantiated in the mind of each person. The experience, or some part of it, must be associated to representations that are already shared for the message content to be shared between two people.
If you are clued into what melody a person is tapping, it is easier to "hear" the melody in the taps. A tap is an example of passing data between people. When computers pass data, they are already programmed to interpret that data in specific ways. Humans do not need to be programmed, but we do need to share some representations for minimal amounts of meaning and data to be shared.
In fact, data requires meaning to come first in human beings. To recognize something as data, is meaningful itself. If you do not understand something is data, is a pattern, then you cannot see that it would have meaning. The converse is also true, if we find patterns, we tend to go looking for the meaning in them, even if there is none. In human beings is that meaning proceeds data. Because it to apprehend that some data is a message requires the apprehension that the content of experience is data with meaning.