Two ways of knowing

Can today computers "know" things? There is a sense in which we say they are "mere" machines, another sense in which they do calculations, solve puzzles, analyze databases and so on. In many areas they do this "calculation" much better than we can, at least billions of times faster and more precisely. And this is just the beginning.

However many people feel reluctant to attribute "reasoning" to computers. After all, they don't seem to be aware, conscious, of the data they are treating. We feed them electrical signals, they treat them in a useful fashion and then give out an electrical signal which we then interpret. The meaning, the significance of the patterns they absorb and they put out only have meaning for an "I", which the computer does not seem to have.

This difference between automatic calculation and real knowledge or wisdom, one performable by a machine, the other demanding an "I", seems to be rooted in our own experience of the way we calculate and know. In fact we do not know whether a computer is aware or not, we can't even know if any person is aware or not, the only thing we can be sure is that each of us, in this precise moment, is conscious. Anything else we believe. What we do know for sure is that there are many processes within our minds that are "calculations" in the sense that we attribute them to computers.

For instance, finding the keys in the keyboard. I do not even have to look at my keyboard, I just think of what I want to say and the words appear on the screen. My hands work "automatically". They already "know" where the different keys are, and they do everything with me having to think about it. Another example is equilibrium. If I had to think of the relative weight of each of my members and then try to imagine possible configurations in which all of that weight would stand on just two small feet, I would have to wait hours to perform a single movement. Yet, I can ride a bike with no problems. In a few milliseconds all of my body react to the constant changes that are needed for my balance when I speedly ride through mountain trails. Our mind generally works as a vast machinery, well programmed for the daily needs, and, most of the time, we just put it on automatic pilot and let it role.

So why can't computers be the same? We program them like we program our brains, and then they do their things "automatically", just as we do. Programming is of course the difficult part. Learning how to walk can take many months even for someone as intelligent as a child who can learn a new language faster than any adult. But after the programming has been done, for walking, language, reading, playing the piano or writing with a keyboard, all the rest is "automatic".

The parallels between the brain and the computer seem so obvious (even in their shortcomings - difficulty with dealing with novelty), that it is difficult to avoid saying that the computers are "mere" machines, just like our brain, most of the time.

But there is another way of knowing. What we described so far is the correct relation between inputs and outputs. A function that works. This way of knowing, in its simplest form, is just to say "if you have x then you should provide z", or "x leads to z". Of course the function may be extraordinarily complex, like face or language recognition. But whatever the level of complexity may be, what characterizes this sort of "calculation" is that it will be correct when we can be sure that the answer given is the appropriate one.

A completely different kind of knowledge is the one we get from listening to instrumental music. One can be changed by listening to music, but what does music say exactly. A Mozart melody, the Vangelis theme that promenently appeared on Carl Sagan's Cosmos, or the Wagner theme that opened the 2001 movie. All these melodies seem to teach us something, they inspire us, they can change the way we live, the way we look at the world around us. But what exactly is understanding a melody?

Certainly it is not the mere ability to recognize it again. It is like seeing a color. When you see yellow you are not only gaining the ability to recognize that color again, you are having an enriching experience that a description of frequencies, amplitudes and polarities would give you. I would like to say that this is knowledge by direct contact. This is the kind of knowledge that no machine can have.