I am a studying Natural Language Processing and came across multiple types of ambiguity like lexical, scope, global, attachment and coordination ambiguities.
Lexical ambiguity: Ambiguity of a single word, like book, study ‘There was not a single man at the party’, two interpretations can be, there was not a single person at the party and there was not a bachelor at the party This ambiguity can be removed using parts of speech tagger(lexical disambiguation).
Syntactic ambiguity(structural ambiguity):
There are 2 types Scope ambiguity and attachment ambiguity
Scope ambiguity: operators and quantifiers can introduce ambiguity. Old men and women went to the church, does it mean (old men and women) or ((old men) and women)
Attachment ambiguity: This occurs when a phrase can fit different points in a parse tree thus creating ambiguous meanings The man saw the girl with the telescope, it is unsure if man with the telescope saw the girl or the man saw the girl with the telescope
Semantic ambiguity: When meaning of words can get ambiguous, then that type of ambiguity is semantic ambiguity. X loves her mother and Y loves too The car hit the pole while it was moving
Anaphoric ambiguity: Anaphoras are entities that are previously introduced into the discourse. Horse is running up the hill. It is very steep. It is getting tired. Anaphoric reference of ‘it’ can cause ambiguity.
Pragmatic ambiguity: the toughest to handle in nlp. This requires a clear understanding of context, when there is ambiguity in context, its pragmatic ambiguity. Usually information inferring capacity is low.
I would like to know, how does a human interpret sentences and identify ambiguities? Are there any materials that help in understanding and interpreting them?
Thank you


There is some research to answer the "how", I am mainly aware of a community called Computational Neurolinguistics where they do experiments with humans and language understanding and relate the results to artificial intelligence techniques. You find some pointers in https://dl.acm.org/citation.cfm?id=1625183

However, there is no clear understanding of "how", we just know that certain tasks are harder and/or can be impaired (damaged) independently from other tasks in natural language processing. For example it is possible to more or less lose the ability to understand language while having a sense of grammar (Wernicke aphasia), on the other hand it is possible to lose the ability to form grammatical sentences while that person can still understand some part of the meaning of sentences (broca aphasia). See https://en.wikipedia.org/wiki/Aphasia

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.