We will now take a closer look at the NLP way of analyzing texts.
Any communication or speech act is built from seven distinct processes: intention, generation, synthesis, and perception, analysis, disambiguation and incorporation. Of these, the first three take place in the speaker or sender, and the last four happen in the hearer's mind. In this context, only generation, analysis and disambiguation are of interest to us.

During generation, the speaker makes a choice of words or symbols appropriate to what he wants to convey to the hearer.

Analysis means that the perceived string is being processed by the hearer in order to extract the possible meanings. This consists of both syntactic interpretation (also called parsing) and semantic interpretation, taking into account the words' meaning as well as their meaning in the current situation. The result of the analysis of a syntactically correct sentence is something equivalent to a parse tree(words connected to phrases).

Disambiguation, finally, picks out the meaning that has most likely been intended by the sender, as some syntactically correct constructs allow for more than one semantic interpretation. You cannot know exactly what the sender wanted to express without having direct access to his knowledge.

On this page: 3.1 - The Grammar of Formal Languages | 3.2 - Parsing|

The Grammar of Formal Languages

Certain rules apply to the structure of a message (a series of symbols with a special meaning). Of course, the message must be formulated in a language common to both the sender and the hearer; this can be either a formal - invented - or natural language - like English, German or Chinese. Natural languages can be partly represented by special formal languages. The basic parts of such a language are terminal symbols - symbols (tokens) that are final, i.e. words in natural languages. These terminal symbols form phrases - parts of speech that stand for certain grammatical categories like nouns, verbs etc. Noun phrases, for example, describe nouns in detail - "the red herring", "the one I saw" etc., while verb phrases express action, behavior or state - "is dead", "reads quickly" and so on. We will refer to noun phrases by NP and to verb phrases by VP. These and other categories all combine to create a complete sentence S. These groups are called nonterminal symbols. In this context a language is a set of strings composed from terminal symbols according to a series of rules, the grammar. The terminal symbols (the words) must be present in a lexicon - a list that is subdivided into sections for nouns, verbs, adjectives etc. that includes all allowed words for this language. The grammar itself provides the framework for building sentences from phrases which in turn are built from terminal symbols. The single grammar entries are called rewrite rules because they substitute a part of the first string to obtain another. They have a similar form to this sample rule S - NP VP which expresses the fact that to build a full-fledged sentence S you need a noun phrase NP and a verb phrase VP (or vice versa: you can combine a NP and a VP to obtain a S). Of course there are analogous rewrite rules for NP, VP and all other phrase types. There may also be several rewrite rules with the same left-hand side symbol; e.g. in addition to the one above there might exist a S - NP VP CON NP VP, where CON stands for a conjunction. This rule could also be expressed as S - S CON S, as any occurrence of S can be replaced by NP VP according to our first rule. <---

Natural Language Processing | Project of Multimedia Systems EECS 579 | update: 22/12/2000 | Daniele Quercia