Structure and classification of free word-groups

The term 'syntactic structure (formula)' implies the description of the order and arrangement of member-words in word-groups as parts of speech. For instance, the syntactic structure of the word-groups a clever man; a red flower may be described as made up of an adjective and a noun. i.e. A + N; of the word-groups to take books, to build houses — as a verb and a noun. i.e. V + N.
The structure of word-groups may also be described in relation to the head-word. For example, the patterns of the verbal groups to rake books, to build houses are to take + N., to build + N. The term ‘syntactic pattern' implies the description of the structure of the word-group in which a given word is used as its head.
According to the syntactic pattern word-groups may be classified into predicative and non-predicative. Predicative word-groups have a syntac tic structure similar to that of a sentence, e.g. he went, John works. All other word-groups are called non-predicative. Non-predicative word- groups may be subdivided into subordinate (e.g. red flower, a man of wisdom) and coordinative (e.g. women and children, do or die).

Structurally, all word-groups can be classified by the criterion of distribution into two extensive classes:
Endocentric word-groups are those that have one central member functionally equivalent to the whole word-group, i.e. the distribution of the whole word-group and the distribution of its central member are identical. For instance, in the word-groups red flower, kind to people, the head-words are the noun flower and the adjective kind correspondingly. According to their central members word-groups may be classified into: nominal groups or phrases (e.g. red flower), adjectival groups (e.g. kind to people), verbal groups (e.g. to speak well), etc.
Exocentric word-groups are those that have no central component and the distribution of the whole word-group is different from either of its members. For instance, the distribution of the word-group side by side is not identical with the distribution of its component-members, i.e. the component-members are not syntactically substitutable for the whole word-group.