ABSTRACT
Phrase structure rules are rewrite rules for constituents of the form A→X1…Xn, e.g. S →NP+VP. This rule should be read as an instruction to replace the sentence symbol S with a noun phrase (NP) and a verb phrase (VP). Thus, the symbol to the left of the arrow is replaced by the symbols to the right of the arrow. Parentheses are used to denote optional elements, and curly brackets are used for alternatives. Phrase structure rules can also be read as specifying the relationship of immediate dominance in a well-formed tree diagram. In the above example, S has the symbols NP and VP respectively as immediate constituents. Phrase structure rules are subject to a set of formal constraints: there must always be a single symbol to the left of the arrow whic is replaced by one or more symbols (a chain) to the right of the arrow. It follows that neither the right symbol nor the left symbol may be zero, e.g. neither 0=Adj+N nor S=0 are possible. Also a chain cannot consist of nothing, nor can deletions follow in the derivation. Transpositions are also prohibited; NP+VP→VP+NP is ruled out. These restrictions are necessary to ensure that each phrase structure rule corresponds to a branching in the tree diagram. The duplication of the phrase structure rules by the tree diagram ensures the reconstructability of the derivational process. The basic components of generative grammar are derived from phrase structures rules. PS-rules are usually context-free, i.e. their use is independent of the environment of the symbols. Context-free rules are distinguished from context-sensitive rules, especially in the earlier versions of generative grammar. For example, a contextsensitive phrase structure rule for the verb visit would be V→Vtrans/#Ndu obj: replace a verb by a transitive verb if a direct object noun follows, e.g. Philip visits Caroline. ( also phrase structure grammar)
References
phrase structure grammar
