What is meaning (1)

What is meaning (1)

Based on Kazuhisa Todayama’s Introduction to Philosophy. What is Meaning?

In this book, the answer to the question, “Is there such a thing as meaning, what is meaning, and in what way?” begins with a question that goes one step further than the frame problem, “How can we build a robot or computer that understands meaning?

Thinking about such a robot can be expected to lead to the fact that if a robot, which is only a thing that obeys physical laws, can realize “understanding of meaning,” it will give us a hint to answer the question of how meaning can be written in a world of only things.

As an example of a robot, let’s consider an appliance robot that acts in its environment according to commands. For example, imagine that if you ask it to clean a room, it will do so, and if you ask it to bring you beer from the refrigerator, it will bring you beer. At that time, if this robot correctly recognizes “room” or “beer” in the command and works, it can be interpreted as understanding what the symbol means, and this robot can be seen as understanding some kind of meaning.

If there is a symbol and it is clearly mapped to something that exists in reality, then we can assume that the “meaning” of the symbol is guaranteed. One way of thinking about this is to abstractly interpret that “existing in reality = the logical value of the represented symbol is true,” and if the logical value of the represented symbol is true, then the symbol has meaning. If the logical value of the represented symbol is true, the symbol is considered to be meaningful. Using this logic, it is possible to guarantee that all symbols are meaningful. This approach is the very approach of formal semantics that I mentioned earlier.

Since the above approach is very simple and can be abstracted mathematically, various theoretical extensions such as many-worlds semantics are being considered. The discussion so far has been about what it means for a symbol to have meaning. This is a different issue from the question of whether robots really understand the meaning. This is another issue.

Here, we will take the approach of symbols having meaning further and apply it to abstract symbols. For example, an apple and a strawberry share the abstract concept of “red”, a ball and an apple share the abstract concept of “round”, and these are further abstracted into “color” and “shape”.

As shown in the figure above, each symbol can be thought of as being connected to each other by some kind of relationship. These relationships can be mathematically represented (abstracted) as functions (see “Structures and Algorithms”), and can be formally manipulated (proved/inferred). In addition, as an extension of the aforementioned discussion, if the relationship is true, it is meaningful; if it is false, it is meaningless, and it is easy to infer that this truth or falsity (meaningfulness) can also be formally manipulated.

Toda-Yama further develops the “meaning” that comes from the connection of these symbols and discusses the “understanding of meaning. Specifically, he further analyzes this relationship and argues that there are two types of semantics: Casual Semantics, which is based on causality, the basic interaction between things in the natural world (X means A ↔︎A, and A alone is the cause of X), and Objectivist Semantics, which is based on the idea that the relationship between symbols is the cause of something. ), and Objectivist Semantics (the relationships between symbols are defined by some purpose).

The idea that the root of “purpose” in Objective Semantics comes from the will of the organism to survive, and that this leads to the problem of the mind to understand, seems to be a direction to the problem of will and mind. For example, if such an algorithm is introduced to AI, it may be possible to understand intentions.

From the perspective of an engineer, I would like to go back to the formal approach of the relationship between symbols and find the answer.