Levels of abstraction

While I was having lunch with “the guys” I began talking about “how we know what we know.” One friend interjected that what I was saying was “too abstract” to be of interest. This has happened to me before, and in such social situations switching topics is fine with me. My friend then went on to say that his interests are “practical” and he prefers to talk about real-world problems and possible solutions.

In a sense, “practical” conversation seem to lack an abstract quality, but in truth all words and all thoughts are abstractions, even words or thoughts about practical subjects. We can use words like “chair” in discussing practical matters, like preferring to sit in a chair instead of the floor, but in that case, the use of the word “chair” is abstract. In such case, we are not speaking about a particular chair, but the concept of “chair.” All concepts are abstract by nature, and lacking this abstract quality our ability to speak in general or broad terms would be impossible.

When we move to the particular and begin to speak about a specific chair, we essentially abandon concept and instead rely on sensory experience. Sensory experience does not require concept; a burning ember feels hot even if we are not familiar with the concept of “hot.” So it is for all direct sensory experience; ice feels cold, noises are soft or loud, textures are rough or smooth, seen objects have form, and so forth.

Accordingly, from birth onwards our sensory experience of the particular evolves into an ability for abstraction; as we accumulate memories of direct perceptions, concepts develop into which we then place patterns we perceive. “Chair” as concept includes an unlimited number of variations, depending upon the conditions of patterning. For some, a “chair” requires a pattern of four legs, a seat and a back; for others no back is required and perhaps just three legs. Conceptual abstraction is flexible and non-particular.

The animals figures on the cave walls at Lascaux, France, are conceptual abstractions; bison, antelope, and horse drawings created 35,000 years ago attest to innate human powers of abstraction. It may be that animals other than humans have the ability to form abstract concepts as well; we certainly treat some of them, our pets in particular, like they do.

As computer scientists work towards developing Artificial Intelligence, their great stumbling block is abstraction. Computers excel in finding patterns at great speed; programmed algorithms are simply formulas designed to examine data and recognize patterns. What algorithms cannot do, as of yet, is generate abstract thought. Computers recognize only a literal reality, the reality of the particular, even if that is ten-billion particularly complex bits of information. Someday perhaps, when posed with a problem to solve and the computer asks “why?”, we will have crossed over into the realm of Artificial Intelligence.

We cannot simply be practical or we would be like stupid machines, unable to form meaningful conceptual hypotheses about ourselves, others or the world-at-large. Human practicality requires the abstraction of language, and the myriad ethical, moral, cultural and personal complexities that accompany abstract concepts.

This does not mean we must talk about “how we know what we know” while eating a hamburger at lunch, but recognize it or not, beneath the appearance of “practical” discussion lurks the unavoidable abstractions of conscious and unconscious mind.



Leave a Reply

Your email address will not be published. Required fields are marked *