Linguistic inspiraton for Commonsense

Consider this sentence - John gave a ball to Jack. John -- gave a ball to -- Jack. This association between John and Jack has an explicit connector - “gave a ball to”. All sentences connect 2 or more entities with a set of words present between them as the link between them. But what about the words which immediately succeed each other? There is no word in between them. The connector between them is the commonsense in the story conveyed by the sentence. Consider the successive word-pairs - John --?-- gave. The Connector is : ‘with his hands’. The Connector comes between John and gave or rather connects John and gave. John (the body) and the physical act of giving (which is “physically further” to it) have this link between them. Gave --?-- ball. Connector : ‘supported by his palms’. The Connector connects giving and the ball by being the link between them. The physical act of giving (the jerking of the arms from John towards Jack) and the ball have this between them (the supporting palms). Ball --?-- to. Connector : ‘moved/transferred’. Moved comes as a connector between ball and the ‘to’wardness toward Jack. ‘The ball’ and ‘the ‘sense of it being directed to someone’ contain ‘it moved’ between them. To --?--Jack. Connector : ‘into Jack’s hands’. Jack's hands come between the pointing direction of ‘towardness’ and ‘Jack’ - the body - which is further to it, as the connector. That - 1) John gave it with his hands, 2) the ball was supported by his palms 3) the ball moved/trans-located 4) the ball went into Jack’s hands - are the commonsense associated with this sentence, deriving from a Linguistic “source” as shown above.

Commonsense in Language-Understanding, at the most fundamental level

Consider this sentence - A gave a ball to B. You understand from the sentence that 'a ball' was the thing given. That is, if someone asks you - what (thing) was given? or A gave what (thing)? - you would answer "a ball". Now the reason you understand that 'a ball' was given is not because 'a ball' immediately follows 'gave', since if that was the case, then in the sentence 'A gave to B a ball', you would say that 'to B' was the thing given since 'to B' immediately follows 'gave'. This leads to an important theory as regards to the coming together of two words/chunks (here, 'gave' and 'a ball'). Whenever words follow each other, their isolated meanings need to be combined in ways so as to make the combined sense match with a PLAUSIBLE REAL-WORLD STRUCTURE. And this is where commonsense (i.e. in making it match with a plausible real-world structure) comes into play. So, when 'gave' is juxtaposed with 'a ball', what do we have to do? Firstly, what is 'gave'? Well, 'gave' is 'gave' - (the act of) 'giving'. What is 'a ball'? 'A ball' is simply 'a (one) ball (that exists)' i.e '(there is) a ball'. Now the mind has to think how to combine these two meanings (word-semantics) to make a resultant sense which matches with a plausible real-world structure. How can you combine '(the act of) giving' and '(there is/exists) a ball', (in that order)? One and the only obvious way that comes to mind is that (the act of) giving was done of the ball (that exists) i.e. 'the giving of the ball' was done. This requires commonsense knowledge since you have to know of "a ball (or a thing) being given" in general. Hence the meaning of the juxtaposed chunks - 'gave' and 'a ball' - is taken (understood) as that "a ball was the thing given".

Commonsense thinking instincts

Consider this sentence - A gave a ball to B. Commonsense says that – now the ball is no longer with A. This is not explicitly stated in the sentence but something that occurs to you. How? You compare two states – A having given the ball to B (which the sentence makes you understand), with an earlier state – A about to give the ball i.e. A having the ball. Why were you inspired to or happen to compare the 2 states? Because you noticed a change in the image of A’s hands – ‘no ball after giving’ and ‘previously having the ball’. How did you notice the change in the image of A’s hands? You saw A’s hands (without a ball) and were REMINDED of a SIMILAR (with a little difference) image of A’s hands having the ball earlier. And why did you notice the change in the image of A’s hands? Well that’s one of the instincts of thinking – noticing changes in things, if and when they occur. Suppose you see a painting on a wall. Then you turn around and after some time turn back to see that the painting is not there, you may/will instinctively remark – where’s the painting gone? You were responsive to or were sensitive in catching a change. Now the only question which remains is – why did you notice A’s hands when the story of the scene is moving towards the ball being with B? Well, you just “looked around”, which is again an instinct of thinking. You would not be able to justify this with a solid reason (as to why did you, after following the ball to B’s hands, looked BACK). If you see some event happening somewhere on the street (say an accident or a crash) you will for some moments also ‘look around’ the car; whereas here there is an even stronger motivation to look around because you are having a look at an earlier-mentioned part of the whole story. If you see a movie changing focus from X to Y over some time, it is very likely that it would occur to you – What happened to X then? This example highlights some of the traits of commonsense thinking – noticing changes, comparing current and previous states of something, being reminded of something similar from something, and lastly going back to a previous state in the first place – all for no apparent preconceived reason, but just on instinct.

Internal subjects and predicates

Why does “gave a ball” make sense and “a ball gave” doesn’t? What does this tell us about language? The format of subject-predicate doesn’t just limit to the whole sentence. It also extends to internal chunks of a sentence. Lets see how. Firstly, what is a subject and what is a predicate? Subject is what you want to talk about, whereas predicate is what you want to talk about what you want to talk about. Now it is a fundamental structure of language that first we mention the former – what we want to talk about – and then the latter. This appeals to commonsense also. IF we mentioned the predicate first and subject later, the listener wouldn’t know what is being talked about, till the end when the subject would be mentioned. That would be odd for cognition purposes. Consider the sentence – John gave a ball to Jack. Here John is what is being talked about and the predicate is – gave a ball to Jack. But even internal to this sentence, look at the chunk – “gave a ball”. Why do we say “gave a ball” and not “a ball gave”? Because we begin with the subject, which is “giving” or “something what given” and the follow it up with the predicate – talking about the “giving” – which here happens to be ‘what was given’ i.e. a ball. Hence every other chunk in a sentence is a pair of “subject-predicate”, if we take this term in its broad sense. The very left-to-right progressive order of a sentence is inherently driven by the ‘subject-predicate’ pairs of chunks of words. Just as another example, “gave a ball to Jack” is the subject-predicate pair of ‘gave a ball’ and ‘to Jack’ respectively.

"Neat" sentences v/s "fuzzy" sentences -

"Neat" sentences v/s "fuzzy" sentences - This is a short write-up. This is just to classify sentences into 2 kinds – Those whose semantic story can be built from their elements (words) step-by-step, and most often also with a clear-cut picture for each element, and those which cannot be built so. Consider this sentence - Peter is holding a basket in his hand. Here the whole story can be built part-by-part by attaching one new element (word) to the previous and that too with a clear picture for each. Peter is a clear entity. Then comes Peter's hand (attached to him). Then comes a basket. And then comes Peter's hand, on the handle of the basket. Now consider this - The two cars crashed against each other. This story cannot be built step-by-step from the constituent words of the sentence. You can begin by saying - there are 2 cars. But then the accident is understood by the cumulative gelled semantic effect of "crashed against each other". This cannot be split as {crashed + against + each + other} in a step by step semantic-cognition process. If at all one has to construct a story like the first-type above, of this sentence, then one might say - There is a car. Its front came to a point. There is another car. It came from the opposite side. It came with its front to the same point. Thus there was a collision of the 2 cars.

COMMONSENSE DEFAULT EMPHASIS -

Every sentence contains multiple items, which make up the total information to be conveyed to the listener. But some item is “critical”. Critical because that is the end-point, or the bottomline amongst the information that the speaker wants to convey. For example, when someone says the sentence – ‘water bodies on maps are shown in blue’, it appeals to us via commonsense that the most important thing which he wants to convey is ‘BLUE’. That’s the endpoint emphasis. Of course he wants to talk ABOUT water bodies and talk WITH REFERENCE TO maps. But somehow it is clear to us that the intended point of the sentence is ‘blue’ – that they are shown in blue. Consider another sentence – John is the culprit. Here it appears that people were trying to identify who is the culprit and this information came from somewhere/someone that John is the one. So the intended point of information in this sentence is ‘JOHN’. This point is also important in the cognitive regard because more often than not, this intended point is the point from where the thought occurred/began in the mind which later became the expressed given sentence. ‘John’ struck someone (in the thought) as the answer to the mystery of who is the culprit, and he later expressed the sentence – John is the culprit. This is the commonsense story behind the sentence. Another point here is that we involuntarily imagine contexts from sentences. This imagined context is principally inspired from this intended point of information. For example, if you hear -John sat on the office chair for the first time, the intended point of information that you automatically sense is ‘the first time’ (and very close in importance comes the part ‘office chair’) which makes up build up a background/context that John is appointed to some new position/post and today is the beginning. So we see three concepts about a sentence, intertwined here – the intended point of information, which is related to (inspires) the imagined context of the occurrence of the speech by the listener and the inception of the thought in the speaker’s mind which was later converted into the spoken sentence.

Why do we make commonsense assumptions?

Consider this sentence: Jack opened the door. Whenever we are given a sentence we make commonsense assumptions about it - things that are not explicitly stated in the sentence. But why do we make them? A part of what makes a door a door is its surroundings. (Just like the meaning of a sentence resides partly in the surrounding text). A sentence tells a story. To imagine anything about a story, we need to imagine the individual elements with something in each one’s surroundings. The way the words of a sentence fuse with each other (to create meaning) contributes to these surroundings of these individual elements. The above sentence has a door. So that’s just a rectangular wooden piece which is a door. But “opening the door” makes it mandatory to imagine a frame or a wall around it since you cannot “open a door” which is just an isolated wooden rectangle standing on the ground. Hence come the other assumptions like John was wearing clothes, that he opened it with his hands, he was looking at the door while opening it etc. Another point- sometimes what happens is that we imagine a context for the sentence. This happens because once all the risings of the words have been done or the sentence has been fused fully, there is a need for a surrounding for the sentence as a whole, which manifests as the context. In other words, the earlier part refers to what can be called internal or sub-contexts whereas the latter can be referred to as external context.

English-to-reality-adjustment = commonsense

Consider the English sentence - John got wet in the rains. Consider the fragment - ‘John got wet’. It implies ‘water on John’s body’. How do we make this cognition? Water is a new word on the implied side (not present on the left hand side), and so is body. How does the “expression” ‘John got wet’ BECOME ‘Water on John’s body’. There is so much fitting in and adjustment of commonsense concepts here. Firstly, here are the chunks - 1) John = A male human being. 2) Got wet = covered with water. 3) In the rains = inside the rain streams. Now, English-to-Reality adjustment = commonsense. Chunks (1+ 2) --- includes the commonsense---> 1) “covered with water” means some physical body covered with water 2) John taken as John’s body. Chunks (2 + 3) ----includes the commonsense---> 1) Rain is water 2) water wets. Chunk (3) ---includes the commonsense---> surrounded on and around by rain streams (among other possible interpretations like say, “John being inside the rain streams/drops”). So real experience of all these concepts is also critical, especially being really under/in the rains to understand the phrase “in the rains” as “under and surrounded by rain streams”.

The “real” medium of commonsense

Suppose your friend told you - I gave a book to John. You will make commonsense implications from the above sentence, like for example - which book did he give to John? Whose (author) book was it? etc. But what made you believe or assume in the first place that the book was a standard book (one with a title, author, price, publisher etc.)? It is the whole medium that you are in, experiencing that reality (the data) which puts you in a certain frame of mind of making fundamental assumptions about everything that you see in that medium. It can be very indirect. Because you know a real, typical human being (your friend) gave a book to another real, typical human being that you subconsciously believed that the item given was also a real, typical, commonsense-satisfying, standard entity i.e. the book. You were convinced of the standardness of the experience (friend telling you about his friend) and that effect spilled over onto the whole data (including the book), i.e. you assumed the book to be a standard one from the surrounding standard scenario of two real, typical human beings transferring something between them. Per se there is no need of believing that the book was a standard one (it could have been one without a publisher or without any cost or whatever - say a memoir of Marvin Minsky prepard by his students and distributed in the department). But because you are “in that medium of ‘reality’ which you are convinced of being so in by the mental registry of the fact that you are actually listening to some real person narrating something to you in person, and where you know that thus commonsense applies, that you apply that state of mind to the book also. Using commonsense is something like “being immersed in the medium, at various root-levels”.

Commonsense

What is commonsense? A gave a ball to B. Commonsense tells us that A gave it with his hands into B's hands. This comes from ball, actually from 'gave ball' But that raises the point that what is the status of 'gave ball' per se? That is what is 'gave ball' perceived as before the "hands come to the mind"? Because gave ball has to be understood as something, to convert it into or bring into mind the hands! I think it is just a tiny video clip of a slight jerk going away (to whatever in the surroundings). Such a clip because the symbolic commonality in all givings is that something "moves/goes away from the giver, to the taker". That what I think is the first impression of the linguistic entity - 'gave'. So the first impression of 'gave a ball' is that a "ball (perceived as what a ball is like) goes away". When the sentence is completely heard, the fact that A and B are 2 people in the scene gets noticed (after the B at the end is heard). In round two, there is a attempted rearrangment of the consumed entities to construct the first mechanical model (who's there? whats where? what goes to/becomes what?) of the scene. It is after this, that is that the entities have fallen in place in some mould, that the above commonsense should flash from "person, giving, a (small) object, to another person" collectively. What does this tell us about commonsense? Even commonsense knowledge asssumption over a given data is a "reasoning process". The fact that a small object was given, the fact that a person gave it and the fact that another person was given it, all contribute to the assumption/imagination of the so called commonsense knowledge. Whats the difference between commonsense knowledge and commonsense thinking? Firstly, call it commonsense reasoning or not, there is a thinking process to build the / to make occur the commonsense knowledge assumption. And then later there is obviously commonsense reasoning to weave commonsense assumptions/facts into something.

Numbers

Numbers don't distinguish between entities. Consider 5 - apples, houses, cars etc. What is in a number that I can have 5 items of anything? Or what property is common to an apple, house, car, box etc. that there can be something the same, irrespective of which of these items it is? This has to be something cognitive. Firstly, this has to be about boundaries of entities. Boundaries is where sameness ends and difference begins. It is on the basis of this sameness existing (till a boundary, after which difference begins) that we are able to call anything, a thing, in the first place! This is the cognitive meaning of a ‘THING’. This concept has extended metaphorically to abstract things like class, strength, Math, cricket etc. wherein the sameness is in some or the other regard (than physical appearance). So the world is made up of only “same” and “different”. ‘Sameness or Difference’ is the only distinction that can exist between any 2 things in the world.

Basis of commonsense knowledge assumptions

What is commonsense? A gave a ball to B. Who has the ball now? B. If you are going to someone’s house, you expect there to be air in the hall (and rooms). You don't expect nitrous oxide or carbon monoxide. This is a commonsense assumption you make. This is very very plausible. But what is the basis of this? Statistics (all houses have air in them)? Or Experience (all houses you have been to have had air in them)? Or sheer “Worldliness” (things being the same in this world, across places)? Variables are samples of constants. Constants are common across entities, variables change. A cricketer is a constant, Sachin Tendulkar is a variable. How do you know which is a constant across scenarios? There are two levels of perceiving data - looking at the constant (general) and looking at the variable (specific). And these are levels - the same thing can be a general or a specific according to which level you are looking at the data at. A bag with some name (say, rotary club) written on it is general (bag) as well as a specific (a carrier example). Rotary club bag is a specific (of the bag) as well as a generality of different designs of rotary club bags. The word you use to represent something is critical. That determines its level. Empirically speaking, the natural default form of perceiving and referring to data is ‘general’. So when I enter a gym for the first time I perceive the body builder guiding people, for the first time, as a trainer (general) and then as say Alex or Jim (specific). Across gyms the trainer (general) is a constant, whereas the Alexes and the Jims vary. So when I enter any new gym I expect my constants to be there as they are and the variables to be different. That is, when I enter any gym I expect there to be trainers but not necessarily Alex and Jim. When I enter a house I expect there to be air around (and not nitrous oxide or carbon monoxide) but not the same shape of the dining table as is in say my house. This is the basis of commonsense knowledge assumptions - why and how we make them or consider them as plausible.

What is ‘meaning’?

Meaning is synonymous with understanding - you “understand the meaning of” something. John is driving a Mercedes. Level 1 - Language : Definitions of the words and mechanical connection of those of the words with each other, according to the rules of English. A male human being is driving (definition of driving - is seated on the front seat with his hands on the steering and his legs on the controls below and the car is moving ahead) a car - Mercedes. Level 2 - Commonsense : Strict - the tyres of the Mercedes are rotating. John is not blind. Plausible assumptions - John is headed somewhere. John is wearing clothes. Level 3 - Deeper Analysis : Something specific about the Mercedes car. (This requires domain knowledge). Say, the battery is discharging at a rate of X.

COMMONSENSICAL CONTRADICTIONS

Whenever you assert any truth, you disassert some other “truths”. Empirical law : ‘A’ and “not A” don't coexist in commonsense. Actually they may. (Note : The ‘A’ is a predicate/property of some entity/ies.) What is NOT A? Every kind of a thing, or a thing for that matter, has a distinct identity because there is a set of properties in it which collectively aren't present in ANY OTHER thing. So in a sense, everything is unique! In fact, if it was not unique (in this sense) it wouldn't be a thing in the very first place. ‘Not A’ is meant in the above sense. John is a doctor. So John is not an architect. Architect is NOT a doctor. However, a person can be practicing both these professions. But as far as commonsense is concerned, a person who is a doctor is not an architect. John is in a saloon. So he is not in a hospital. Hospital is NOT a saloon. In reality, there might be a saloon inside a hospital. A and B are enemies. So it is only commonsense which tells me that A does not serve freshly baked cookies to B. because this gesture is expressing love and care which is NOT trying to harm each other which is a basic property of enmity. Although it is a possibility that their enmity is such that it allows leeway for such gestures. A commonsense contradiction is less exact/strict and more far-reaching in conclusions, than a logical/scientific/mathematical contradiction. In short, saying “the phone is to my right” implies logically only that it is not to my left, but commonsensically (along with the same) also that it is not to my northeast. If that was the case (northeast), it would be misleading the informant.

COMMONSENSE BELIEFS -

Most Aussies are fighters. John is an Aussie. So we believe that he is a fighter. John and Jack are friends. Friends know what they are doing in their lives (like studentship, job or business or whatever). So we believe that John and Jack know what the other is doing in his life. This works because knowing what the other does is BECAUSE OF the fact that they are friends. However, at the same time, that doesn't mean that all friends will know about each other. Note : this belief system works, it doesn't mean it's necessarily always true. But if I say this road is accident-prone (because there have been many accidents on this road), then it is wrong to be scared of the “road” per se because the cause might be a building on the road from whose terrace some miscreants who live in that building throw stones at cars. So the road is not per se the cause of the accidents. However, at the same time, the generic fact remains the same that you should be scared of that road till you don't figure this out since from the human mind’s perspective, correlations are the precursors to arriving at / detecting causes. If most parents love their children we say parents love their children. Commonsense fact/belief : A --- B most of the time. If a is a sample of a, then a---B. This is true only if ‘A’ness is the cause of A----B. That is, the link A----B should be true causally because of A. So if I say (most) Aussies are fighters. Then believing that a given Aussie is a fighter is right only if “being an Aussie” is the cause of the fighting spirit of the Aussies.

COMMONSENSE KNOWLEDGE MACHINERY IN UNDERSTANDING INFORMATION

John gave a ball to Jack. What all have I understood when I say I have understood this sentence? One thing I understand is - Before this (John giving the ball to Jack) happened, there were John and Jack existing/being. Then, John gave the ball to Jack. How do I say that John and Jack existed before? If you are doing something, then you are alive. If you are alive now, you were alive just before too. This gets processed by my commonsense. The fact that I process that John and Jack existed before is evidenced by the fact that one way in which I can explain the meaning of this sentence is - “There is a John. There is a Jack. And the former gave a ball to Jack” (like a story). The first two sentences are the evidence. Why did I create a story like this? One reason - Temporal sequence of events is always a good way to understand something. So basically by using that commonsense what I have done is a temporal dissection of the sentence. Another reason - it's always good to lay down the facts before stating the connections/relations amongst them. That helps comprehension. What was the mechanism of the commonsense processing? 1) I want a temporal sequence. That directly relates to ‘context’ - context is always “before in time”. Context explains the background scenario or the backdrop. John and Jack existing before is the backdrop / background scenario and thus the context for understanding this sentence. So we see that one direction for commonsense machinery to work in is to extract the context of the given information. Another thing I understand is - John gave the ball with his hands to Jack. How did this commonsense processing work? The presence of ‘ball’ in the part “gave a ball”. Balls are generally held in and handed over (given, in this case) via hands into the other person’s hands. Why did this process work? On being told that a ball was given to Jack by John, a followup logical story-completion question asking for more details, (which is a trait of understanding), which arose in the mind was - how was it given? Here is a basic link-list of - information and typical follow-up questions - which express the nature of our commonsense understanding process of anything given to us. Thing - what is it? Person - who is he? Process/action - how? Relation/connection - why? These get processed by default in the mind upon coming across information-elements. Answers to these questions are by commonsense if the information is commonplace, otherwise by asking the question to the source. So the 2 directions in which commonsense knowledge machinery works on given information, are towards context (for general comprehension requirements) and “picture-completion” (for specialized follow-up details)

LANGUAGE AND PERCEPTION - COGNITIVE DEFINITION OF A SENTENCE

Sentences describe. What do they describe? Sentences are combinations/fusions of descriptions of aspects of things/phenomena. John was riding a bicycle. John is a thing. His aspect, namely, ‘past action’ is described by ‘riding a bicycle. Riding is a thing. Its aspect ‘beholder’/’agent’ is described by ‘bicycle’. The sentence is a combination of the above two. The ball is in the dustbin. Ball is a thing. Its aspect - location - is described by ‘in the dustbin’. I was singing while cooking. ‘I’ is the thing. Its aspect - action - is described by ‘singing’. Singing is a phenomenon. Its aspect - timing - is described by ‘while cooking’. The sentence is a combination of the above two. One can notice that this aspect that is being talked about is the hidden element, not present in the sentence as or associated with a word. This is the cognitive component of the sentence - the thought about something that arises in the mind while wanting to express something. THe mind wants to talk about the location of the ball and out pops it - “The ball is in the dustbin”. You don't use the word standing for the aspect - location.

ROLE OF LITERALNESS (IN VISUALIZATION) IN SENTENCE-UNDERSTANDING

Consider this sentence - John was filling water into the jar. Now, it occurs to us that John must be filling water FROM SOMETHING into the jar. The key here (as discussed before) is visualization. Even if you go by the possibility-gauging procedure discussed before, you still spot the lacuna at the level of the visualization only - of a connection/link (John ---filling---> water). The point is that there is also a LITERAL construction step of the components of the sentence that happens. It is only when you try to connect ‘John’, ‘filling’ and ‘water’, by trying to visualize “John filling (with his hands ( which comes with the definition of filling)) water” LITERALLY, that you realize that the water needs to be held by something. (And this very specific realization is obviously because of the commonsense knowledge of gravity/support that gets invoked in the aforesaid mental process - literal visualization). So now we have so far discussed 3 mental conceptual elements of sentence-understanding (they might be overlapping each other) - Visualization Literalness Possibility-gauging Question about a different topic - ‘commonsense’ - arising from the above discussion : Consider this sentence - John was riding a bicycle. Here we commonsensically understand and assume that obviously John was riding the bicycle on the ground (or some base). But this doesn't “occur to us”, like it occurs to us that John must be filling water FROM SOMETHING into the jar (when given the earlier sentence) even though the basis for that too is commonsense. Why so? My guess is that in the earlier sentence, there was “a jar being filled with water into”. So there was a need for a “‘complement’ to the jar, from where the water was filled” for the completion of the story. Also the bicycle scenario is much more commonplace in terms of riding a bicycle ON A GROUND/BASE than that of filling water into a jar FROM SOMETHING. The basis of the invocation of the complement to the jar, as well as that of the mental-picture-building of the ‘ground’ were both commonsense. But the mental processes in which commonsense was invoked were different. This tells us something basic although noteworthy about commonsense-knowledge-invoking. It is that there is “specific/instantial” intervention of commonsense in some cases where the scenario is unique/specific, versus that in more general scenarios in others. The piece of commonsense is the same - gravity - but the scenarios in which it got invoked were opposite. In the first case, it “occurs to us”, in the latter it gets passively assumed.

LEVELS OF UNDERSTANDING DATA - A THEORETICAL IDEA

Suppose someone tells you or you read somewhere - Person X gave a ball to person Y. Now, the definition of a ball is - a spherical object used in playing games. But when I am listening to or reading the sentence at the very very first sight, what is my mind’s idea of a ball? I take just a tiny aspect of the definition of a ball. I don't process it as “a spherical object used in playing games” at that time. I am not “in touch with” the actual definition of a ball while merely reading the sentence. I just faintly, quickly, hazily visualize a typical ball, which is my “understanding of the read word ‘ball’ and move ahead. I similarly do the same visualization for “some person X” (just say a blurred image of a vertical human being), “some person Y” (just say a faint image of a vertical human being), and “giving” (which is a 1 second video clip of hand(s) jerking ahead). This is LEVEL 1 - understanding the sentence at a mere Syntactic English level or making sense of just the English of the sentence. LEVEL 2 - If I then reflect, think, pay closer attention to or begin to ponder over the data, I will draw leads and inferences and clues from it. One thing you might say is - “were they playing a game or something?” LEVEL 2 immediately follows LEVEL 1 in time. In fact they are over ‘chunks of words’ of a sentence at a time, if the sentence is not short like the example above. Meanings of words are like Graphical User Interface operating system files presented in the mind as icons with a symbol and a name. The name of the file is the very word. The icons (symbols on top of the words) are the representative faint, hazy visualizations of the “meaning” of the word. The mind accesses just these symbols during the first-sight reading of or decoding the mere English (syntax) of the sentence. Thinking or reflecting( LEVEL 2 - understanding the Semantics) is when you “double-click” on the icons of the files and “open” those files which have their definitions/properties stored inside them. In LEVEL 2 mode, (after the basic syntax is congnized), you access the contents of the files, and also inter-connect the contents of the different files to draw inferences/leads/clues from the sentence. So when you open the file ‘ball’, one thing you immediately see (say, written at the very “beginning” of the file) is “..playing games..”, which makes you pass the remark - “were X and Y playing or going to play some game?” (Basically, was this related in any way to playing some game?).

POSSIBILITY OF THE SEMANTICS OF A SENTENCE

Reference - The knowledge representation scheme above. Consider this sentence - John killed the bird with a gun. We think about the HOW / POSSIBILITY when given a sentence (i.e. the “way” in which it is possible, if so). In the above case that is - There was John. There was a bird. And John fired a bullet from it to hit the bird. The bird died. This is how what the sentence says is possible. (The other non commonsensical way in which this can be true is that John took a bird in his control and hammered a gun on it repeatedly to kill it. But commonsense is not what we are discussing here. We are only discussing possibilities). This possibility-gauging is a primal component of understanding any sentence. How do we judge this possibility, given a sentence? First step - We form a literal mechanical connective model of the English components of the sentence. (Ref the knowledge representation scheme before). The second step is checking the above with our experiences of the real world i.e. we check the validity of 1) the connections and 2) the properties (of the entities and of the connections). There can be invalidity at 3 levels - 1) one node of a connection is wrong 2) both nodes of a connection are wrong 3) a property is wrong. *The connection itself cannot be wrong by itself unless its a word out of the dictionary. Examples - One node is wrong : E.g. The dead man fired a bullet at the police. The connection here is ‘fired bullet’ between the ‘dead man’ and the ‘police’. Clearly the node dead man is wrong. Both the nodes are wrong - e.g. the dead man killed another dead man. The dead man cannot kill and a dead man cannot be killed. So both the nodes are wrong. Property is wrong - The blue phosphorus reacted with oxygen. The property ‘blue’ of phosphorus is wrong. There is nothing like blue phosphorus. Invalidity in possibility-gauging is almost equivalent to not understanding something, (the other case of not understanding being if you don't know the meaning of a word or a phrase).

COMMONSENSE IN CAUSALITY where human intent is involved -

If X is the reason and its said resultant responsive action is Y, you still cannot be sure if the commonsense function of action Y is the reason to address the reason X. So if X is the reason and Y is its said action in response, it is only commonsense assumption in the first place that the intention of the commonsense function of Y is to address the reason X. Illustration : Consider this - I am feeling hot. So I switched on the AC. Reason - I am feeling hot Responsive Action - I switched on the AC. The reason why the action was done could be that I like hearing the sound of AC while sweat is pouring down my body. So the reason the reason is the reason for the action may not be the commonsense function of the action. It is only a commonsense assumption that the cooling effect of the AC was desired to neutralize the hot feeling.

THE ALL-KNOWLEDGE PROJECT - PURE FUN!

Take a humongous source of data. All the possible kinds of sentences (on the basis of their parts-of-speech-combinatorial structure) will be covered in it. For each distinct type of sentence-structure do this : Say, the structure is NVN - John was doing his homework. Make combinations of all possible Ns and Vs in it from the dictionary. You will have a set of all possible sentences in the English language of the structure - John is doing his homework. Similarly generate all the possible sentences from all the various different structures of sentences. You will thus have all the possible grammatically correct sentences in the English language. Some of them will make sense. Some won't. If you distribute this set on the internet for kids to play with and just say whether this sentence is “possible” or “not possible” (which in fact will be a commonsense test), you will have a set of all possible sentential English language knowledge.

LANGUAGE AND VISION : VISUALIZATION IS THE ONLY DIFFERENTIATOR IN MEANING

Consider these 2 sentences - Bill drives a truck. Bill stays at Campbell. Are these 2 sentences saying the same thing or different? Answer - Different. Why? Firstly it is clearly naive to say that if the words in 2 sentences are different, then those 2 sentences mean different things, since that is clearly not true. One would say that they “MEAN” different things - their meanings are different. How do you know their meanings are different? What is the meaning of the meaning of theirs in the first place? Also, you cannot do an algebra of it by expanding the constituent words into their definitions and then cancel the same words on both the sides (i.e. in the 2 sentences) and see if nothing remains on either side. One would say that the “stories” conveyed, or the relationship between the constituent entities conveyed by each sentence (which is what meaning broadly is) are different. One says that there is a man Bill and there is a truck and Bill drives that truck. The other says that there is a man Bill and a place called Campbell and Bill stays at that place called Campbell. But then this is recursive. How do you know that the meanings of “there is a man Bill and there is a truck and Bill drives that truck” and “there is a man Bill and a place called Campbell and Bill stays at that place called Campbell” are different? They may mean the same thing. The only way out of this is our visual machinery. The visualizations of these 2 sentences or of the respective sets of words in quotes just above LOOK different. It is equivalent to seeing 2 different pictures. Why is a computer mouse different from a pen? Because they mean different things. Why? Because they LOOK different or the visualizations of the functions (if you want to differentiate them by their functions) of each - “controlling the pointer on the screen” and “writing on something with a hand” - LOOK different. What makes “different” different? Why is a Compaq mouse different from a Logitech mouse which looks exactly the same? Because one is manufactured by Compaq and the other is manufactured by Logitech. Why is “manufactured by Compaq” not the same in meaning as “manufactured by Logitech”? Because they are different companies i.e. those 2 companies LOOK different from each other, or at least are at different places from each other (which makes their visualizations distinct from each other). That’s the only differentiator in meaning - VISUALIZATION.

WHY IS THE ‘VERB’ THE “KING” IN LANGUAGE?

Consider the following 3 sentences : John is playing the guitar. John was playing the guitar. John will play the guitar. These are the same things happening at different points in time. The separating or differentiating aspect of these 3 sentences is only time. So why does that differentiator have to be Linguistically manifested via the form of the verb only? Why not the noun or adjective or whatever? John exists today, existed yesterday and will exist tomorrow. The guitar exists today, existed yesterday and will exist tomorrow. The ‘playing’ (of the guitar) ALSO exists today, existed yesterday and will exist tomorrow. So why does the temporal differentiator have to be plugged into the ‘playing’ (verb/action) only? (when in fact all the 3 words - John (noun), guitar (noun) and playing (noun) are equivalent with respect to the difference in the time). Why not say that ‘John-ed plays guitar’ for ‘John was playing the guitar’, the “ed” indicating “John past”? Why is it so that whatever (that has an action/verb) that happens, becomes an “event/episode that happens i.e. verb”? (That is, say, why not “event-ed happen”?) Why does the verb carry the representative onus of the event (when there are other parts of speech hanging around)? Language is verb-centric.

CONTINUITY IN SOLIDS -

Suppose I lift a TV from the top, commonsense says that the base of the TV would also move upwards. What is the explanation of such inferences of continuity? If a thing “looks” different from others, it is a different thing.If there is continuous sameness in any respect then it is the same thing till the point that that sameness is obeyed. In fact, this is how we define a ‘THING’ in the first place. For calling anything in the world a “thing”, there has to be sameness, in some respect, throughout that thing (i.e. throughout all the constituents of that thing). This is how we point to anything, or talk about anything as the subject. Coming to specifically the physical things, we do not perceive the physical world around us as, say, some one distinct object and an imaginary section of some of its surroundings as clubbed together with it to form one thing. The moment something is identified to be a ‘thing’ (as defined above), the effect of any action on any constituent of that thing is perceived to be on that whole thing. (When I stab a point on the surface of a table with a compass, it is perceived to be stabbing the ‘table surface’ or the ‘table’, and not just a point on the surface (where the tip of the compass actually hits the table surface). This explains - when we push, say, a TV remote it is commonsense that the remote moves as a whole, and not so that just the localized portion of application of force moves or gets displaced. In other words, this is talking about constraint relations in Mechanics. So when I lift up a TV from the top, it is commonsense that, say, the base of the TV would also move up.

Fundamental property of a sentence

EVERY string of words (collection of successive words) from the beginning of any sentence, BEARS/HAS AS A PROPERTY, the next adjacent word. Lets see this in the case of some examples - 1) John has a book. John + has : John has the property of having. (John what? John has). (John has) + a : John's having/possession has the property of oneness. (John has what? John has one 'something'). (John has a) + book : John's possession of oneness has the property of book. (John has one what? John has one book). 2) John has trouble working. John + has : John has the property of having. (John what? John has). (John has) + trouble : John's having has the property of (the possession of) trouble. (John has what? John has trouble) (John has trouble) + working : John's having trouble has as a property the attribute of working. (John has trouble with what? John has trouble with working). 3) Pictures of John have great value. Pictures + of : Pictures have the property of being owned/represented by. (Pictures what? Pictures of something) (Pictures of) + John : Ownership of pictures has the property of John as its beholder. (Pictures of whom? Pictures of John). (Pictures of John) + have : Ownership of pictures by John has the property of possessing/having. (Pictures of John what? Pictures of John have something) (Pictures of John have) + great : Ownership of pictures by John has the property of possessing/having greatness. (Pictures of John have what? Pictures of John have greatness). (Pictures of John have a great) + value : Ownership of pictures by JOhn has the property of possessing/having greatness which has the property of being the attribute of value. (Pictures of John have greatness of what? Pictures of John have greatness of value). 4) Experience has been unkind to John. "Experience has" : The experience (which John has had) bears the property of having something (unkindness). "Experience has been" : The having of the experience bears the property of being. "Experience has been unkind" : The being of the having of the experience bears the property of unkindness. "Experience has been unkind to" : The unkindness of the being of the having of the experience, bears the property of being towards something (John) "Experience has been unkind to John" : The towardness of the being of the unkindness of the being of the having of the experience, bears the property of John.

THE FUNDAMENTAL RELATIONSHIP BETWEEN LANGUAGE AND THOUGHT -

Given 2 (or more) pieces of data or thoughts (thoughts are ultimately, data), ADD them. Given some data or thought (say, X), delete a part of X (say, x) from it. But for doing this you should first see X as being composed of x and (X-x) for removing x. So another way of putting this point is that given (A+B), remove B from it. E.g. - I don't want this tree in this painting. So rephrasing the above - Given A and B, do A+B Given (A+B), remove B The above principle applies even in the case you want to combine A and some part/aspect of B to say, draw an inference or to create something or whatever. Taking an aspect of B is (2) and adding it to A is (1). It also similarly applies if you transform A to A’ since that is only repeated subtraction and addition with itself. So any package of thinking can be explained in terms of (as a collection of) the above 2 points. You cannot do anything with just A i.e. just with one piece of data. (Typically, you have some knowledge stored in your mind to combine it with and draw an inference from A).

FUNDAMENTAL COGNITIVE COMMONSENSE ALGORITHMS FOR GOALS -

1) I want X (I know of the state beforehand of being possible) -> What do I need for that? 2) I don't want X (Here you want the “opposite” of something. Here you don't QUITE have a perfect or an exact idea of what state you want. The goal is to just eliminate/undo the present state) -> What do I need to get rid of for that? Note : If at any conclusion-state the conclusion is “readily” available/possible in that state, ACT/EXECUTE IT. (This is the barest minimum commonsense). Then continue the above process. Example for Type 2 - My desk is unclean. Goal : I don't want my desk to be unclean. (Type 2). Autoresponse : What do I need to get rid of for that? Ans : Dirt. Conclusion : So I want to get rid of the dirt. This again becomes type 1. Autoresponse : What do I need for that? Ans : A cloth/rag. Conclusion : So I need a cloth. This is type 1. Autoresponse : What do I need for a cloth? Ans : I need to go to where it is kept. Conclusion : So I need to go to where it is kept. This is type 1. Autoresponse : What do I need to go to where it is kept? Ans : I need to walk. Conclusion : So I need to walk. If this is readily available (unless say, you are surrounded by a mess of furniture touching you on all sides preventing you from moving. Or say your legs are injured), ACT. Note that this function is in fact checked at every conclusion-state above. It fails at every earlier conclusion state (you cannot readily get rid of the dirt (i.e. it’s not going to disappear on its own), you cannot readily get a cloth (it's not going to fly and pop up into your hands) - if it does it (say, someone throws it at just that moment) ACT and continue the ‘process’, and you are not automatically going to land at the place where the cloth is kept). After you reach the place where the cloth is kept, the above process repeats. Also note that all the ‘Ans’ at every stage at retrieval of bare commonsense knowledge, which is behaving like a maximum-squeezed-version of the chain of commonsense thinking. Example for type 1 - I am somewhere nearby a building complex. Goal : I want to enter the building complex. Type 1. (I know that building complexes can be entered. This is the state known beforehand being possible). Autoresponse : What do I need for that? Ans - Gate/entrance. (This comes from the commonsense knowledge piece : A building complex has a gate to enter from.) Conclusion : Look for the gate/entrance.

WHAT IS THE MEANING OF ‘MEANING’?

Of Word, Sentence & Text. WORD - Suppose I say - John likes cricket. What is cricket? It is a game. Is this sufficient? No. Why? Because you haven't mentioned what is (/what happens) in that game? So one property of the meaning of the word X is to explain ‘what does X “contain”?’ Now let’s see what this “contain” contains. What would you mention about something to explain the meaning of that thing? Physical object external form, function/use/purpose. Abstract entity/concept class, Relation between constituent entities/concepts. Process Conditions under which it happens, dynamics, results/effects. Description Manner or degree of a property, effect SENTENCE : Relationship between 2 or more words. TEXT : Relationship between 2 or more sentences. Semi-illustration (sentence) - ‘John likes cricket.’ John - Human being (external form) + who does various things (function) Likes - derives happiness from it (condition) + and hence does repeatedly (result) Cricket - game (class) + in which two teams compete with a bat and ball in such and such way (relation between constituent entities/concepts). Meaning of the sentence - Which constituents of the respective words are entangled to compose the meaning of the sentence? 3 versions : Human being + derives pleasure from and does repeatedly + engage in a game. (External form) + (condition and result) + (class) The sentence talks about the conditions and results of a certain form engaging in a certain class of thing

What is the meaning of INSIDE ????

What is our cognitive undestanding of containment? Closed from all sides. What is our mind’s commonsense understanding of closedness, and thus of insideness? Firstly, there is something particular about all solids - all things. It is that anything that is solid is full-dimensional (which happens to be 3 dimensions) in its existence. That is, its surface covers all “sides” POSSIBLE, locally. Consider a foot-ball. You see its front curved surface facing you; then you rotate it, say, to the right. Then you happen to see that whatever was remaining, which was not seen earlier in the front view, is closed. The right side now is closed because your memory tells you that it is the same thing as what you saw upfront in the beginning. So the football is closed throughout. And closedness throughout leads to the concept of insideness. Anything which has a closed surface on ALL POSSIBLE sides is said to be inside that surface (3D), or loop (2D). So the key check for insideness is the ‘boundary’ or the ‘border’, ON ALL POSSIBLE SIDES.

A new language

Consider this funny sentence : John sandwichized. Suppose adding 'ized' to anything meant in our present-day language - hold it in your hand, put into your mouth, chew, swallow. Suppose there was no word like 'ate' (or 'eat') at all taught to you. And by the above, we understood what we TRULY UNDERSTAND by "John ate a sandwich." In a sentence, there is an indicator (word/chunk of the sentence/part of a word) for every concept. Learning language is learning which indicator stands for which concept in reality. For example, the suffix 'ized' to sandwich here makes us understand that a sandwich was eaten. The trouble with our language is that - there is a PARTIAL word “ate” for the literal process of eating, standing for - 'putting in mouth + chewing + swallowing'. Hence something is left to commonsense (which here is that ‘ate with “hand”’). If our language was like the way its suggested above, there would be no need for commonsense! This vaguely inspires something like - what if we have a language in which there are words, for all different collections of words which form the "commonsense package" associated with the present-day-language partial words, in various scenarios?

Psychology v/s Logic

Suppose the given data is - David is teaching vectors to Bryan. And they are in the same class (same age). One commonsense inference - So David is better at vectors than Bryan. The question is - why does a particular inference strike us with more likelihood when we come across a piece of data? Here, there are other inferences like - 1) So Bryan is learning. 2) So Bryan knows (would know) what a vector is. 3) Davis seems to be a helpful student. etc. to name a few. The machine and humans would know the piece of commonsense fact that ‘Teachers are generally older than their students’. This would contradict line no.2 of the given data. And in the LIGHT OF THIS CONTRADICTION, this very inference should be inspired - i.e. So David is better at vectors than Bryan. (They would now be compared in their abilities since they are now known to be fellow-students). This wouldn't have been the primary inference if it wasn't given in the data that they are of the same age. This implies there is a motivational fluid of impending conclusions running through, while the mind is doing commonsense reasoning. That is why some inferences get a preference over the others i.e. humans are more likely to jump to them than others. Logic cannot distinguish between these (unless of course these internal motivational rules themselves are somehow fed in as additional logic-statements into the machine).