In the last post, I made a brief mention of Nasruddin, Heuristics and Non-Linear Thinking. I shall continue in this post on the same theme and establish the role of semantics in Design and Non-Linear Thinking, and explore some techniques of practicing non-linear thinking.
First, here is a very brief synthesis of the previous posts on this subject: In the previous posts, I applied law of 2 and tried to demonstrate the underlying unity of the theme in various disciplines – which is referred to as Plausible and Demonstrative, Creative and Logical, Synthesis and Analysis, Right Brain and Left Brain, Linear and Non-Linear, Heuristic and Rational and so forth. Largely, the conclusion that one can reach is that there are two distinctive types of thinking – which are interdependent on each other, but at the same time have very different functions and applications. We termed the functions of these two types as – recognition of problems and finding solutions to problems. We also presented a few laws of synthesis – from various fields – particularly societies, management, computing science and consciousness.
Before we get into Semantics, let’s examine linear thinking and non-linear thinking in some detail – with some examples and techniques of how to switch from one to another.
Logic – at least the way Socrates and his disciples developed it – is largely linear. This is carried over into mathematics as well. Logic mostly depends on reasoning from evidence. The early Artificial Intelligence (which is neither artificial nor intelligence) applications ran into some problems with the linear reasoning and there are now several attempts to develop non-linear reasoning systems. Such logics in AI are called “non-classical” reasoning systems. Here is a classic example of linear and non-linear reasoning – it involves a very well known logic puzzle.
Three monks and three cannibals were traveling together because of some strange circumstances – even though naturally there cannot be any friendship and trust among such a diametrically opposite groups. If at any time, the cannibals outnumber the monks – they will eat them. Now, they came to a river bank and they had to cross the river. There was only one boat. The boat can only carry two people at a time. Now, your problem is to devise a strategy to transport the monks and the cannibals safely across the river.
The usual linear reasoning system will start with River Bank-1: 3M, 3C, River Bank2: 0M, 0C as the initial state of the system, and will try to devise a series of moves that will establish the desired final state - River Bank1:- 0M, 0C, River Bank2: 3M, 3C. It is a simple puzzle – you start with something like first two cannibals will travel, one will come back, two monks will then go etc., and you can easily arrive at the final state.
Suppose you give the problem to Nasruddin, what do you think he will do? He will listen to your descriptions, and even before you have completed your narration, he will jump from his chair and shout with the excitement of a child “I know, I know - the monks will take the bridge, and the cannibals will take the boat”.
It is a perfectly valid solution to the problem – isn’t it? Then you tell him “look Nasruddin, I did not say that there was a bridge”. Nasruddin will reply, “Well – you did not say that there was no bridge either”. Now, you modify the problem description, and you will add another constraint to the problem – there are no bridges. “Well, in that case, the monks take the helicopter” will be Nasruddin’s answer. By this time, you get very frustrated and shout back at him – “I want you to solve the problem – not avoid it”. But, for Nasruddin, there was no problem there to be solved. Can you see how he eliminates the problem instead of trying to solve it?
No matter, how hard you try, you cannot contain Nasruddin. He will always think of exploiting some constraint that is not included in the problem definition and use that as a means to eliminate the problem.
The AI community in the last thirty years tried and discovered several techniques for “programming Nasruddin type behavior” into the computer programmers – with some reasonable success. I have no intention of boring you with the Non-Monotonic Reasoning systems here. But what is the technique of Nasruddin?
One, Nasruddin does not depend on “evidence” as the only way of establishing the existence of something. All formal logical systems depend on evidence as the only means of establishing “truth”. For example, if I tell you that all visitors from Canopus are fools, and you know that Mr. X is a visitor from Canopus, you can confidently establish the truth that this particular Mr. X is a fool. But, there is a catch. Suppose, if this particular Mr. X is not a visitor from Canopus, then what are you going to do? You cannot establish anything. He may or may not be a fool. You can conclude Mr. X’s foolishness, only if you know that he is a visitor from Canopus.
We need facts to recognize truth, but Nasruddin does not rely on such petty things like facts to recognize truth. According to Nasruddin, some thing does not have to be a fact for it to be true. Kurt Gödel proved this mathematically – it is known as Gödel’s incompleteness theorem. The incompleteness theorem states that in any formal system, there always exists a statement which is “known” to be true, but cannot be proved.
Rabindranath Tagore – with the brevity that a poet alone can command – says it beautifully.
“If I say that the earth is flat – you inform me that my near sight is false.
If say that the stars are the fireflies attracted to the moon, the candle of the sky – you inform me that my far sight is false. My dear logician – in your presence, I prefer to be blind”.
In the specific case of monks and cannibals – all Nasruddin does is to bring his knowledge of the larger environment to solve the problem. The minute you tell him about the river and the boat, he can think of a bridge. He can bring his knowledge of traveling in the air, or swim in the water. Or, he will say that the monks – using their magical techniques – will vanish and reappear on the other side. Or, he will say that the monks will use their secret herbs and put the cannibals to sleep. His possibilities to come up a solution are infinite.
The study of system level relationships requires knowledge of semantics – or the “meaning” of things. There are basically two types of relationships – structural and semantic. It is important to distinguish between these two types of relationships. Though, in logic and other sciences, we are mostly taught how to study the structural relationships, in fact, our mind works naturally very well with semantic relationships.
A few examples from various fields may be in order here.
In language, grammar defines the structural relationships. In English, a valid sentence must have a subject, a verb, an orbject/predicate. And, there are various rules that define the relationships between various parts of speech. These relationships are basically structural – they govern the structure of the sentence. We can say that water is triangular – it is a grammatically valid sentence, but completely meaningless.
In information systems there are basically four types of structural relationships. By information systems I do not mean just software applications. Any system that organizes knowledge and depends on that organization is an information system. By this definition, societies, cultures, various disciplines of study are all information systems.
The first – and simplest structural relationship is hierarchy. This relationship in information systems is called “IS-A” relationship. A “IS-A” B. This could be interpreted in several different ways:
• B is the parent of A
• B is the boss of A
• B is part of A
• B is subsumed by A
• B is the root of A
Entire systems, organizations, societies are modeled based on this one particular relationship. Most hierarchical structures use this relationship as the primary relationship. The different sections in this blog-page have a parent-child relationship with the master page. There are basically five sub-sections – header, footer, posts, author information, archives – each of them is a “child” of the main blog page. Similarly, many corporations have a hierarchical structure – with the owner of the corporation at the top. The “semantics” of this relationship connotes ownership and protection.
The second structural relationship is network relationship. In such a structure, elements, or objects at the same level can be connected. This is called “sibling” relationship. A “IS-A brother of” B. The relationship basically means that two things are at the same level. It connotes competition and/or collobaration.
There are two other types of structural relationships in information systems hyper-text and groups. Both these relationships are somewhat non-linear relationships, but nonetheless, they are structural relationships.
The Set, and Hyper-Text are more complex structural relationships than a simple hierarchy and a Network. A set uses a function to create a structure. For example, you can say that all integers belong to a set. You can then define a function that determines an integer and therefore, the membership to the set. The members in the set may not have any particular relationship – they have something in common – that’s all. This is how most social groups are formed. The function gives an identity to the set and therefore to all its members. All people “born” in US are “Americans”. “Being born” in the US determines the identity of the set called Americans.
A Hyper-Text relationship is much more interesting relationships. It does not relate two nodes, but instead it relates content from one node to another node. All web pages use this structure extensively. For example, I can provide a link to an article in wikipedia in one of my articles in this blog. This does not mean that this blog as a whole and wikipedia share any relationship; it also does not mean that this article and the wikipedia share any relationship, and it also does not mean that this article as a whole has any relationship to that article in wikipedia. This relationship connotes friendship. There is no function, there is no responsibility, and there is no ownership in this relationship.
Using these four relationships, we can explain almost all “order” in our world. We defined the basic family relationships (parent-child, sibling), groups and friendships.
So, where is the place for Semantic Relationships?
We conveniently left out the most important, most complex, wonderful, problematic and the most beautiful relationship of all– that of the lover and the beloved. None of the above relationships explain that relationship – do they? What is the relationship of a husband and wife? What is the relationship of a teacher and a student? Friendship comes closest, but these relationships transcend friendship.
This is the realm of semantics – which is the topic of next post.