Christopher Manning > Quotes

“Most commonly, noun phrases are arguments of verbs. The arguments of verbs can be described at various levels. One can classify the arguments via semantic roles. The agent of an action is the person or thing that is doing something, the patient is the person or thing that is having something done to it, and other roles like instrument and goal describe yet other classes of semantic relationships. Alternatively, one can describe the syntactic possibilities for arguments in terms of grammatical relations. All English verbs take a subject, which is the noun phrase that appears before the verb. Many verbs take an object noun phrase, which normally appears immediately after the verb. Pronouns are in the subject case when they are subjects of a verb, and in the object case when they are objects of a verb.”
― Christopher Manning, Foundations of Statistical Natural Language Processing

“As an example, suppose we want to translate a text from English to French. The noisy channel model for translation assumes that the true text is in French, but that, unfortunately, when it was transmitted to us, it went through a noisy communication channel and came out as English. So the word cow we see in the text was really vache, garbled by the noisy channel to cow. All we need to do in order to translate is to recover the original French – or to decode the English to get the French.”
― Christopher Manning, Foundations of Statistical Natural Language Processing

“In his book Human Behavior and the Principle of Least Effort, Zipf argues that he has found a unifying principle, the Principle of Least Effort, which underlies essentially the entire human condition (the book even includes some questionable remarks on human sexuality!). The Principle of Least Effort argues that people will act so as to minimize their probable average rate of work (i.e., not only to minimize the work that they would have to do immediately, but taking due consideration of future work that might result from doing work poorly in the short term).”
― Christopher Manning, Foundations of Statistical Natural Language Processing