Cognitive Approach of Human-Machine Communication

1. Introduction

During the time, computers became an important part of our lives. They have known a great diversity of applications, from solving problems or specific tasks to entertainment and communication, being present in almost all fields of our activity.

Computers can be viewed as systems that transform the input information into output one based on mathematical and logical operations. We interact with them using different equipments and interaction techniques. The devices are specialized in realizing specific input and output function.

The man and the computer are being part of a process of communication that has developed a lot during the last few years. Still the humans perceive the reality and the environment they live in, in a very complex and diverse way and when the peripheral devices and interfaces were designed the complexity of humans' capture, storage and elaboration of information was not taken in consideration. The main peripheral devices remain almost the same: the keyboard and the mouse for the inputs and the display for the outputs, so the improvement is not very considerable. Also appeared different paradigms of indirect interaction based on objects (real or virtual) that make difficult the work with specialized software and limit the productivity.

The aim of the paper is to summarize the tendencies and the research in this field and based on it to propose a new vision of human-computer interaction which will facilitate the work with software and will increase productivity and maybe in the future work will develop new peripheral devices and interfaces.

First of all, all theories should be focused on some premises: a human-machine interface should facilitate the communication between the two entities. A good communication from the human perspective must be as natural as possible, efficient, comfortable and with few or no errors. It is also good to be intuitive and not to need very much training.

2. The Interaction Metaphor

From the cognitive psychology point of view [7], the user can be looked like a cognitive system that is designing efficiently a product, by using the computer.

So the communication between the man and the computer represents actually the communication between two systems: the human cognitive system and the computer system. The two systems interact based on a bidirectional transfer of information from one part to another.

The information is processed and the results can be reflected in the behaviour of the user or of the computer and instantiated in different decisional actions (for example the computer displays a window with a specific image, the man pushes a specific key or set of keys from the keyboard or moves the mouse etc.).

All information present at a given time in the cognitive system (Figure 2) represents the knowledge of the system. The knowledge changes dynamically through processes of accommodation and actualization. Neither the humans nor the computers can access all the knowledge in a given moment of communication, but only a part of it (Figure 1, 2).

Fig. 1. The distribution of computer knowledge via its types of memories [9]

Fig. 2. The human knowledge in the cognitive system

We "know" and "understand" reality only by acting upon it and thus acquiring "knowledge". The knowledge is neither a copy nor a mirror of reality, but the form and content of it are constructed by the human who experiences it [2]. "Knowing" means having the appropriate knowledge. "Understanding" means being able to associate the respective knowledge with other knowledge and take decisions. Humans communicate with other humans based on a mutual degree of "understanding" which is very effective at least for the basic aspects of the human life. While the "common understanding" between humans decrease for more complex issues, human to human communication failure increase.

In contrast with humans, computers are not able to construct their knowledge on their own or to "understand" it. Therefore, the human could not base its communication on a "common understanding" hypothesis, being permanently obliged to know and take into consideration the machine behaviour in a certain context. As the machine complexity grows, the human-machine communication highly charges the human cognition in order to maintain an "understanding" of the machine context. It is thus clear that creating machine capable of "understanding" is of great interest for the improvement of the future human-machine interaction.

To achieve this, we need to accurately identify and define what is "understanding" and how can we create "artificial understanding", that is machines which "understand".

"Understanding" is created by active interaction between the individual and the environment mediated by the cognitive structures of the individual. People think by manipulating internal models of reality. What they learn in interaction with the environment depends upon their own structuring of those experiences. Thus man does not merely respond to the environment, he construes it. A possible solution to construct artificial understanding could be based on what we call the interaction metaphor approach [8].

The interaction metaphors have been already employed to make computers easier for humans to use. The aim of the metaphor is in making a new system look and act like a known system. Metaphors represent archetypes of knowledge that makes possible the explanation of new concepts by the simple usage of them. In this way the content of communication for a certain operation is reduced and the productivity of the system is increased [2].

A metaphor allows us to understand one concept in terms of another, enriching our mental imagery and imbuing concepts with meaningful attributes [4]. As Lakoff and Johnson argue [5], "metaphor is pervasive in everyday life, not just in language but in thought and action. Our conceptual system, in terms of which we think and act, is fundamentally metaphorical in nature." They also argue that the metaphors we hold for abstract concepts provide us a richer way to understand them, allowing us to reason about them and make inferences with them. Metaphor is very often defined as an analogy or a set of analogies with implicit reference.

Probably the most discussed use of metaphor is in software interface design. Common metaphors in use, such as the