Hierarchy in Grammar

In grammar, hierarchy refers to any ordering of units or levels on a scale of size, abstraction, or subordination. Adjective: hierarchical. Also called syntactic hierarchy or morpho-syntactic hierarchy.

The hierarchy of units (from smallest to largest) is conventionally identified as follows:

1. Phoneme
2. Morpheme
3. Word
4. Phrase
5. Clause
6. Sentence
7. Text

Etymology: From the Greek, "rule of the high priest"

Examples and Observations

Charles Barber, Joan C. Beal, and Philip A. Shaw:  Within the sentence itself, there is a hierarchical structure. Take a simple sentence:

(a) The women were wearing white clothes.

This can be divided into two parts, Subject and Predicate, in each of which there is a main part and a subordinate part. The Subject consists of a Noun Phrase ('The women'), in which a noun ('women') is the head, and a determiner ('The') is a modifier. The Predicate has as its head a Verb Phrase ('were wearing') which governs a Noun Phrase ('white clothes') as its Object. The Verb Phrase has a main verb ('wear') + -ing as its head, and an auxiliary ('were') as a subordinate part, while the Noun Phrase has as its head a noun ('clothes'), and an adjective ('white') as a modifier... This notion of hierarchy in sentence structure is of primary importance. For example, if we wish to change a sentence (for example, from a statement to a question, or from an affirmative to a negative form), we cannot do it by rules which just shuffle individual words around: the rules have to recognize the various units of the sentence and the ways in which they are subordinated to one another. For instance, if we want to turn the sentence 'The king is at home' into a question, we have to bring 'is' in front of the whole noun phrase 'the king' to produce 'Is the king at home?' "The is king at home?" would be ungrammatical.

C.B. McCully: Turning to a syntactic hierarchy, we might want to observe that the smallest elements of syntax are morphemes. Whether these morphemes are either nonlexical (as in the plural inflections /s/ or /iz/ -- cats, houses) or lexical (= lexeme -- cat, house), their function is to constitute words; words are gathered into syntactic phrases; phrases are gathered into sentences . . . and beyond the sentence, if we wish our hierarchical theory to account for reading as well as speaking and writing, we could include constituents such as the paragraph. But clearly, morpheme, word, phrase and sentence are again constituents of the syntactic grammar of English.

Charles E. Wright and Barbara Landau: The relationship between semantic and syntactic levels has been actively debated (see, e.g., Foley & van Valin, 1984; Grimshaw, 1990; Jackendoff, 1990). However, one general framework posits linking rules, building on the fact that the semantic and syntactic levels of representation share a similar hierarchical structure: Those thematic roles highest in the thematic hierarchy will be assigned to those structural positions highest in the syntactic hierarchy. For example, in the thematic hierarchy, the role of agent is considered to be 'higher' that either 'patient' or 'theme'; in the grammatical hierarchy, the syntactic function of subject is assumed to be higher than direct object, which is higher than indirect object (see, e.g., Baker, 1988; Grimshaw, 1990; Jackendoff, 1990). Aligning these two hierarchies will have the net result that, if there is an agent to be expressed in the sentence (e.g., using the verb give), that role will be assigned to subject position, with the patient or theme assigned to direct object.

Marina Nespor, Maria Teresa Guasti, and Anne Christophe: In prosodic phonology, it is assumed that, besides a syntactic hierarchy, there is a prosodic hierarchy. The former is concerned with the organization of a sentence into syntactic constituents and the latter with the analysis of a string into phonological constituents. The prosodic hierarchy is built on the basis of the morpho-syntactic hierarchy. Although there is a reliable correlation between the two hierarchies, the correlation is not always perfect (cf. also Chomsky and Halle 1968). A classical example of the mismatch between syntax and prosody is illustrated below:

(12) [This is [[[NP the dog that chased [NP the cat that bit [NP the rat that was running away]]]]]
(13) [This is the dog] [that chased the cat] [that bit the rat] [that . . .

In (12), the bracketing indicates the relevant syntactic constituents, specifically NP's. These constituents do not correspond to the constituents of the prosodic structure of the sentence, which are indicated in (13).