Should Engineers Think Like Computers?

Peter Kreeft is a philosopher at Boston College and author of a textbook on Socratic logic.  So far, so dull, you may think.  I have heard Prof. Kreeft speak on numerous occasions, and "dull" does not apply either to his speaking or his writing.  In the introductory section of his text, he says something that I think engineers need to hear.  It's about ways of thinking and knowing, and the two different competing kinds of logic that are taught today.

If you take a course on logic in college these days, it is most likely going to be what is called "symbolic logic" or "formal logic."  In fact, most electrical engineering undergrads get a sample of this kind of logic when they take digital electronics.  It turns out that digital computers use things called logic gates.  One of the simplest logic gates, called an AND gate, can have two inputs and one output.  As you may know, the numbers in digital computers are in binary form, meaning they have the value of either 1 or 0.  Typically, 1 is represented by a high voltage and 0 by a low voltage.  An AND gate looks at its two inputs, and the only time its output is a high voltage is when both one input AND the other input are high.  Otherwise, its output is low.

           

If you understood that, you have a firm grasp of an element of what is called Boolean algebra, named after the nineteenth-century logician George Boole, who formalized this type of thinking in an attempt to bring the discipline and clarity of mathematics to logical thinking.  Boole's project was taken up by others, and now most philosophers think, do, and teach highly mathematical symbolic logic when they do any kind of logic at all.

But there's a problem with symbolic logic, a problem it shares with computers, whose structure is nothing more than a physical embodiment of symbolic logic.  The old expression "garbage in, garbage out" applies to symbolic logic as well as it does to computers.  Symbolic logic can do marvelously complicated things with its inputs, but it doesn't "know" what the inputs are any more than a copy machine knows Shakespeare after you've copied a page of King Lear with it.  Symbolic logic says nothing about the truth or reality of what you give it.  To understand what things really are, you have to get outside the pristine mathematical structure of symbolic logic and embrace what Prof. Kreeft calls Socratic logic.

It could just as well be called Aristotelian logic or classical logic.  You can get a sense of what gets lost if you think symbolic logic is the only way to think, by reading what Prof. Kreeft says about reason and logic: 

"The very nature of reason itself is understood differently by the new symbolic logic than it was by the traditional Aristotelian logic.  'Reason' used to mean essentially    

'all that distinguishes man from the beasts,' including intuition, understanding, wisdom,    moral conscience, and aesthetic appreciation, as well as calculation.  'Reason' now usually means only the last of these powers." (Socratic Logic, edition 3.1, p. 22)

He goes on to say that as first philosophers and then people in general have accepted these definitions of logic and reason, it's become harder to understand some concepts that earlier ages knew almost without thinking.  Take one of the most basic questions that even five-year-olds can ask:  "What is it?"  That goes to the heart of the being of a thing:  what is a horse or a man essentially?  Not just what is it made of, but what is its true nature?  Symbolic logic doesn't ask such questions.  It is a logic of manipulation and process.  As Kreeft says, instead of trying to understand the fundamental essentials of things, we "think about how we feel about things, about how we can use them, how we see them behave, how they work, how we can change them, or how we can predict and control their behavior by technology."  That last phrase—predicting and controlling behavior by technology—is largely what engineers do. 

And it's a good thing to do, rightly applied, and symbolic logic is essential to the task, especially these days now that everything engineers do involves information technology.  But it is a great mistake to think that symbolic logic is all we need to make our way in the world, and we can safely ignore the questions of what a thing really is on our way to bending it to our technological will.

As Kreeft points out, animals have feelings, use their environment, and even make use of rudimentary tools in some cases.  But we do not share with animals the higher operations of the human mind, including the ability to see connections between apparently different things, and to achieve true understanding about universal concepts such as freedom or love.  No animal and no computer (sorry, all you artificial-intelligence folks) will be able to do that. 

Kreeft is concerned that we will give in to the ancient temptation to idolize the work of our hands:  symbolic logic and the computers it gave rise to.  But if we do so, we will lower ourselves to the level of the beasts, and neglect those powers of the mind recognized by classical logic as what makes the difference between humans on the one hand, and animals and computers on the other hand. 

The takeaway from all this is simply to remind us that scientific thinking in the sense of symbolic logic is not the only kind of thinking or knowing that there is.  Engineers of all people should recognize that they do their work in the context of human society, which is forever beyond the grasp of symbolic logic to analyze or comprehend.  Something more is required.  We can call that something classical logic, or wisdom, or even faith.  But to limit ourselves to the kind of logic that computers can do is to leave our humanity behind. 

Sources:  Peter Kreeft's Socratic Logic, edition 3.1 was published by St. Augustine's Press in 2010. 

A Purely Nominal Problem

 

My father didn’t like to spend money when he didn’t have to, so when my mother expressed a wish for an automatic dishwasher, one day he showed up with an old portable unit that some friends of ours got rid of when they bought a newer model.  It was a big floor-model box on rollers, and you ran one hose to the kitchen sink and another to the sink drain and plugged it into a wall outlet.  It worked fine for a few weeks.  Then one day it refused to drain.  We opened the door and saw all this dirty dishwater, so we bailed it out and I volunteered to fix it.  Because I was cheaper than calling a repairman, my father agreed to let me tear into the thing.  After a lot of gross and messy work, I found the problem:  a toothpick had lodged between the drain pump impeller and the housing.  That little toothpick had jammed the pump, and as a result the whole washer couldn’t drain.

I learned several things from that experience (not the least of which was to avoid appliance repair as a future career).  But the most important one was that fairly small, common, almost unnoticeable things can have big negative effects.  And the things don’t need to be physical ones at all.  In fact, immaterial things can make a lot more difference than any physical object, especially if they are so widespread that you don’t notice them, like fish who don’t realize it’s water that they’re swimming in.  The little thing I’d like to draw your attention to is nominalism.

The word “nominal” is often used by engineers to mean “typical” or “according to the specifications.”  But its original meaning is “relating to names.”  Nominalism is a philosophical position first proposed by William of Ockham (~1288 A. D. - ~1348).  Until he came along, most philosophers thought the word “apple,” for example, referred to a real and essential, though immaterial, “appleness” that is shared by all things properly called apples.  However, William of Ockham claimed that there was no such thing as appleness—the essence of what it is to be an apple.  Instead, “apple” is just a name for certain kinds of objects that we, in our human wisdom, have decided to call apples.  In other words, he denied that there are any universals—that is, essences of things.  There’s just a lot of round red fruits out there that, for convenience, we have decided to group under the name of “apple,” but in reality, all apples are different individuals and there is nothing more to the word than the sum of all things called apples.

After William of Ockham proposed nominalism, the other philosophers had to think of a name to call themselves, and the term they chose was “realists.”  A realist, in this technical sense, thinks that there is indeed a universal concept, objective and independent of our minds, which in English is denoted by the word “apple.”  These concepts, which the moderate realist Aristotle called essences, are as objectively real as a bank account.  A bank account is not a material thing, though there may be material records of it.  A bank account is a non-material concept, and so are the concepts of “apple,” “tree,” “horse,” and “man.”

Unless you are aware of this historical controversy, as a typical 21st-century person you probably think and act as a nominalist most of the time.  For example, if you agree with the words of the 1992 U. S. Supreme Court decision in Planned Parenthood vs. Casey that “At the heart of liberty is the right to define one's own concept of existence, of meaning, of the universe, and of the mystery of human life,” you are a nominalist, because defining is what a nominalist does.  First comes the name, then come the items to be grouped under that name.  But the namer is always in charge, and things can be arbitrarily regrouped by the namer to suit one’s convenience.  As the philosopher Richard Weaver has pointed out, an important consequence of nominalism is that “if words no longer correspond to objective realities, it seems no great wrong to take liberties with words.”  So, for example, the genocide of Jews by Nazi Germany in World War II is euphemized to “the final solution.”

Engineers are perhaps less apt to fall into the grosser errors of nominalism, because we have frequent encounters with objective reality.  If a computer chip you design doesn’t work, calling it by a different name isn’t going to make it start working.  But even the way engineers use logic has been affected by nominalism.  The digital logic that all digital computers use is based on symbolic logic devised by George Boole, a nineteenth-century mathematician whose hope was to reduce all logic to symbols.  The trouble is, symbolic logic assumes that nominalism is true, and throws out a great deal of material that traditional Aristotelian logic relied on, including the notion that understanding is a uniquely human power essential to right thinking.  But if everybody uses nominalist logic that can be expressed by Boole’s “boolean algebra,” we have reduced our thought processes to those that can be done by computers.  This is an important source of the idea put forth by artificial intelligence proponents that the brain is really nothing more than an advanced wet computer.  If we can’t make computers act like humans, we’ll reduce humans to the point that they act like computers.

Let’s hope that doesn’t happen.  Regular readers of this blog may have noted that I try to approach philosophical matters from an Aristotelian perspective that moves from the real, objective world to the world of thoughts.  Nominalism tempts us to do the opposite:  to define things the way we want them to be, and then look for pieces of reality that fit our preconceived notions.  I think engineers of all people should be aware of the dangers of nominalism.  Realism is more than just being practical; it means realizing that there is more to the world than we can possibly understand or control, and the proper attitude toward nature is one of humility.  Otherwise, like that toothpick in the dishwasher, nominalism can throw a whole culture out of whack.

Sources:  The quotation by Richard Weaver is from his 1948 book Ideas Have Consequences (Univ. of Chicago Press), p. 7.  I was inspired to write about nominalism and realism by reading one of the few logic textbooks in print which employ realist Aristotelian logic rather than symbolic logic as its basis:  Peter Kreeft’s Socratic Logic (South Bend, Indiana:  St. Augustine’s Press, 2004).