[an error occurred while processing this directive]



4. The problem of ambiguity

During the interpretation process, we have talked about previously, in which agent B deduces P' from sentence F' (= F) ambiguities of various nature can born. This is due to the fact that the natural language is not a formal language. We are used to distinguish such ambiguities in four classes: lexical ambiguities, syntactic ambiguities, semantic ambiguities and pragmatic ambiguities. Such distinction is not exhaustive but it is useful to focus on the problem and for practical purposes.





Lexical Ambiguity.

We will omit the problem of words written or pronounced not correctly. This kind of errors are simply solvable through the comparison with the expressions contained in a dictionary (like your word process application probably does). What I want to evidence now, is that a same word can assume various meanings. In this case the that ambiguity is generated from the fact that the intelligent agent could not be able to chose the right meant. As an example we consider the "cold" adjective in the following sentences:

" That room is cold. "

" That person is cold "


It turns out obvious that the same " cold " adjective assumes, in the two phrases different meanings. In the first sentence it indicate a temperature, in the second one a particular character of a person.





Syntactic ambiguities.

Do you remember the logical analysis you probably learn at school? Well that one is just the first step. Through the syntactic analysis our program would have to be in a position to isolating subject, verbs and various complements... Sure it can be done, but it is already rather complex. I would not want you to discover yourself "ignorant", but try to open a book and begin to make logical analysis... let me know after how many lines you give up! An example in order to clarify the problem:

" Mario eats the apple. "

Mario: Subject
It eats transitive verb, time present, third singular person
the apple: Object complement.

It look easy... but now:

" Mark, who was not in hungry, asked to Giulio to cook later. "


I do not know about you, but I am not able to solve the problem... and, on the other hand, I would not know how to explain it to my computer! Stop Joking, I mean that developing a good syntactic analyzer is very difficult and however it couldn't completely solve our problems! Let's consider the following phrase as an example:

" I have seen Mario to the university. "


It can be interpreted like: " While I was at the university, I have seen Mario. " or " I have seen Mario, who was to the university. ". This is what we are used to call syntactic ambiguity. It is due to the fact that the sentence has two derivation trees (both syntactically corrected) each one carrying to a different result.





Semantic ambiguities.

To analyze a phrase from the semantic point of view means to give it a meaning. This should let you understand we arrived to a crucial point. Semantic ambiguities born from the fact that generally a computer is not in a position to distinguishing what is logical from what is not. As an example if a friend says to you:

" The car hit the pole while it was moving. "


All of us would surely interpret the phrase like " The car, while moving, hit the pole. ", while nobody would be dreamed to attribute to the sentence the meant "the car hit the pole while the pole was moving ". Why the first interpretation is preferred to the second one? Because we have a model of the world that helps us to distinguish what is logical (or possible) from what is not. In the last example the fact "the car is moving" is logical, while "the pole was moving" is not. To supply to a computer a model of the world is not so easy!





Pragmatic ambiguities.

Pragmatic ambiguities born when the communication happens between two agents who do not share the same context. We imagine to receive a telephone call from a friend far away, even than living in a different continent:

" I will arrive to the airport at 12 o'clock. "


In this case, speaker and who listener, because of various jet lag, does not share the same context and probably one of them will attend to the airport for various hours. I do not know if I have convinced to you, therefore I'm going to show one more example. Suppose someone tells us:

" Outside is very warm. "


The information that we can learn from this phrase is complete only if we know who has pronounced it and his concept of " warmth ". A eskimo, as an example, pronouncing a phrase of the sort could let us know that outside it's 3 or 4 degrees!
This means it is important the hearer knows or could suppose as much as possible about the speaker and his knowledge (say world model).




Italia HyperBanner Network