Syntactic movement is the means by which some theories of syntax address discontinuities. Movement was first postulated by structuralist linguists who expressed it in terms of discontinuous constituents or displacement. Certain constituents appear to have been displaced from the position where they receive important features of interpretation. The concept of movement is controversial; it is associated with so-called transformational or derivational theories of syntax (e.g. transformational grammar, government and binding theory, minimalist program). Representational theories (e.g. head-driven phrase structure grammar, lexical functional grammar, construction grammar, and most dependency grammars), in contrast, reject the notion of movement, often addressing discontinuities in terms of feature passing[definition needed] or persistent structural identities instead.
The a-sentences show canonical word order, and the b-sentences illustrate the result of movement. Bold script marks the expression that is moved, and the blanks mark the positions out of which movement is assumed to have occurred. Each time, movement takes place in order to focus or emphasize the expression in bold. For instance, the constituent which story in the first b-sentence is the object of the transitive verb likes, the canonical position of an object being immediately to the right of the verb. By fronting the object as a wh-expression, it becomes the focus of communication.
The examples above use a blank to mark the position out of which movement is assumed to have occurred. Blanks are just one means of indicating movement, however. Two other means are traces and copies. In transformational grammar, movement has been signaled by a trace t since at least the 1970s proposal by Noam Chomsky, e.g.
Subscripts help indicate the constituent that is assumed to have left a trace in its former position, the position marked by t. The other means of indicating movement is in terms of copies. Movement is actually taken to be a process of copying the same constituent in different positions, deleting the phonological features in all but one case. Italics are used in the following example to indicate a copy that lacks phonological representation:
While there are various nuances associated with each of these means of indicating movement (blanks, traces, copies), for the most part, each convention has the same goal, which is to indicate the presence of a discontinuity.
Within generative grammar, various types of movement have been discerned. Two important distinctions are A-movement vs. A-bar movement and phrasal vs. head movement.
Argument movement (A-movement) displaces a phrase into a position where a fixed grammatical function is assigned, such as in movement of the object to the subject position in passives:
Non-argument movement (A-bar movement or A'-movement), in contrast, displaces a phrase into a position where a fixed grammatical function is not assigned, such as movement of a subject or object NP to a pre-verbal position in interrogatives:
The A- vs. A-bar distinction is a reference to the theoretical status of syntax with respect to the lexicon. The distinction elevates the role of syntax, locating the theory of voice (active vs. passive) almost entirely in syntax (as opposed to in the lexicon). A theory of syntax that locates the active-passive distinction in the lexicon - i.e. the passive is not derived via transformations from the active - will reject the distinction entirely.
A different partition among types of movement is phrasal vs. head movement. Phrasal movement occurs when the head of a phrase moves together with all its dependents in such a manner that the entire phrase moves. Most of the examples above involve phrasal movement. Head movement, in contrast, occurs when just the head of a phrase moves, whereby this head leaves behind its dependents. Subject-auxiliary inversion is a canonical instance of head movement, e.g.
On the assumption that the auxiliaries has and will are the heads of phrases - of IPs (inflection phrases), for instance - the b-sentences are the result of head movement, whereby the auxiliary verbs has and will move leftward without taking with them the rest of the phrase that they head.
The distinction between phrasal movement and head movement relies crucially on the assumption that movement is occurring leftward. An analysis of subject-auxiliary inversion that acknowledges rightward movement can dispense with head movement entirely, e.g.
The analysis shown in these sentences sees the subject pronouns someone and she moving rightward (instead of the auxiliary verbs moving leftward). Since these pronouns lack dependents, i.e. they alone qualify as complete phrases, there would be no reason to assume head movement.
Since it was first proposed, the theory of syntactic movement yielded a new field of research aiming at providing the filters that block certain types of movement, also called locality theory. Locality theory is interested in discerning the islands and barriers to movement. It strives to identify the categories and constellations that block movement from occurring. In other words, one wants to understand why certain attempts at movement fail, e.g.
The b-sentences are now all disallowed due to locality constraints on movement. Adjuncts and subjects are islands that block movement, and left branches in NPs are barriers that prevent pre-noun modifiers from being extracted out of NPs.
Syntactic movement is controversial, especially in light of movement paradoxes. Theories of syntax that posit feature passing reject syntactic movement outright, that is, they reject the notion that a given "moved" constituent ever appears in its "base" position below the surface, i.e. the positions marked by blanks, traces, or copies. Instead, they assume that there is but one level of syntax, whereby all constituents only ever appear in their surface positions - there is no underlying level or derivation. To address discontinuities, they posit that the features of a displaced constituent are passed up and/or down the syntactic hierarchy between that constituent and its governor. The following tree illustrates the feature passing analysis of a wh-discontinuity in a dependency grammar.
The words in red mark the catena (chain of words) that connects the displaced wh-constituent what to its governor eat, the word that licenses its appearance. The assumption is that features (=information) associated with what (e.g. noun, direct object) are passed up and down along the catena marked in red. In this manner, the ability of eat to subcategorize for a direct object NP is acknowledged. By examining the nature of catenae like the one in red here, the locality constraints on discontinuities can be identified.
As mentioned above, movement analyses of discontinuities often posit the existence traces. A trace is what is left behind in a position in which the moved constituent appeared at some deep level of syntax or some point in the derivation before surface syntax. Hence instead of the blank used above to mark the base position of a constituent that is moved, a t (t = trace) is used, e.g.
Traces are considered primarily in Chomskian transformational grammar and its various developments. They are distinguished from other empty syntactic categories, commonly denoted PRO and pro. More details and examples can be found in the article on empty categories.
One way to explain this contrast is to assume that the trace left behind by the extraction of who in the second example blocks the contraction of want and to. Without the trace, this explanation would not be available.
The validity of this argument based on wanna contraction and similar arguments has been called into question by linguists favoring non-transformational approaches. Consider, for instance, that other types of contraction are possible despite the putative presence of a trace, e.g.
In these cases, contraction of the wh-word and auxiliary verb is easily possible despite the putative presence of the trace. What these examples suggest is that wanna is not transformationally derived through contraction, but rather it is a lexical item in its own right that subcategorizes for a bare infinitive. It is, then, similar to other apparent contractions such as aren't and isn't, which upon scrutiny cannot be transformationally derived, but rather must also lexical items in their own right.