X-bar Schema in Generative Syntax: Phrase Structure

A fundamental distinction in syntax is whether heads precede or follow their complements. English and similar languages (German, French) are head-initial, with heads preceding complements. Japanese, Tibetan, Korean, Navajo, Basque, and many other languages are head-final, with heads following complements. This is one of the most common word orders cross-linguistically. German exhibits mixed headedness, with some heads (complementizers, prepositions) being head-initial while others (auxiliaries, verbs) are head-final. Three major movement operations exist: Wh-movement to CP specifier, Head movement to C, and NP-movement to satisfy EPP. Not all languages have all three types—Chaha lacks both wh-movement and NP-movement, while Italian lacks NP-movement.

This section presents a comprehensive movement typology and discusses the status of head movement. There are three criteria: size (head versus XP), landing site (A versus A-bar), and locality. Landing sites are determined by probe feature type. A-bar has an optional feature and can be on a variety of heads within a clause. An A feature is an inherent feature on the DP, so it's not optional. There is a third distinction: limb movement versus spine movement. Things along the spine are the heads of the extended projection and their projections. The specifiers are the limbs. When we look at XP movement, it's always going to be limb movement in English. When we look at head movement, it's spinal movement. The question is whether we ever have spinal A-bar XP movement. The speaker argues that there is such a movement type, coming from work on Nahuatl and Malagasy. VP fronting in English is a bar VP fronting, while VP fronting in Malagasy is this type three movement. Rollup movement is not new to the literature. The speaker introduced C-movement using a Van Irk sort of feature. It's a categorial feature and it's the category feature that's shared by heads in the extended projection from Grimshaw.

This section establishes the theoretical foundation for understanding head movement in generative syntax. It introduces the core principle that every syntactic category—whether lexical (verbs, nouns) or functional (Tense, Agreement)—must contain a head that projects upward through levels of structure. The head-to-head movement concept is defined as when one head relocates from its base position to occupy the head position of another category. The lecture distinguishes between covert movement (invisible to surface word order) and overt movement (visible in word order changes), setting up the analysis framework for examining how heads reorganize syntactic structure to satisfy grammatical requirements like case assignment and feature checking.

Head movement presents a paradox: it exhibits successive cyclic properties (requiring syntactic analysis) yet lacks semantic content (suggesting phonological status). The resolution lies in operation timing: transfer applies before raising, eliminating semantic effects. The head then moves to the edge for subsequent phase processing. Successive cyclic movement continues until reaching a criterial position where halting occurs. XP-YP structures cannot be labeled without shared features, compelling movement. In embedded questions, attempting to raise wh-phrases too far creates uninterpretable structures—embedding yes/no questions within wh-questions produces incoherence that crashes at the interface. This interface-level constraint solves the halting problem, demonstrating how conceptual-intentional requirements constrain syntactic operations.

In syntax, English handles multiple auxiliaries (modals, perfectives, progressives) through VP stacking, where each auxiliary projects its own VP; however, when the normal relationship between tense and the verb is disrupted by negation or question formation, head-to-head movement occurs, causing the tense node to move to the front, which explains phenomena like do-support in questions and negatives.

The Uniformity of Theta Assignment Hypothesis states that identical thematic relationships between items are represented by identical structural relationships at D-structure. Arguments have thematic roles (agent, patient, recipient) that must first map to syntactic positions. In passive voice, the patient moves from its thematic object position to subject position but remains semantically interpreted as the object. This explains why selectional restrictions remain constant despite surface syntactic changes. Major languages consistently map agents to subjects and patients to objects, with causatives adding agents to transitive verbs.

Theta Theory, developed by Noam Chomsky, explains how thematic roles (such as agent, patient, theme, and locative) are assigned to arguments within a sentence's argument structure; the theory establishes two key principles: (1) all entities in a sentence must be assigned a thematic role, and (2) every thematic role must be assigned to exactly one entity, ensuring grammatical completeness and proper semantic interpretation of sentences.

X-bar Theory alone cannot generate grammatical sentences because it does not constrain the number of arguments verbs require. Verbs are classified by argument count: intransitive verbs require one argument (subject/DP), transitive verbs require two arguments (subject and object, both DPs), and ditransitive verbs require three arguments (subject, indirect object, and direct object). The nature of each argument (DP, PP, CP) is crucial for grammaticality. Thematic relations (theta roles) encode selectional restrictions by specifying what kind of entity can fill each argument position: agent (initiator of action), experiencer (one who experiences/perceives), theme (entity that undergoes action), goal (entity toward which motion takes place), recipient (special kind of goal with change of possession), source (entity from which action originates), location (place where action occurs), instrument (object with which action is performed), and beneficiary (entity that benefits from action). There is no one-to-one relationship between thematic relations and arguments; a single argument can have multiple thematic relations. Theta roles are bundles of thematic relations that cluster on one argument.

Theta roles categorize arguments into classes: agents (causers), experiencers (perceivers), beneficiaries (recipients), possessors (holders), patients (affected entities), goals (destination points), and locations. Proto-role theory addresses classification limitations by evaluating which role properties arguments exhibit most strongly. Categories are classified using binary features: [±verbal] and [±nominal], yielding four categories: preposition [-V,-N], noun [+N,-V], verb [-N,+V], adjective [+N,+V]. All phrases are endocentric with heads projecting to phrases. Functional categories (I, C, D) handle finiteness, clause introduction, and definiteness. Binary branching and non-tangling conditions maintain hierarchical structure throughout derivations.

In generative syntax, theta roles are syntactic positions that bundle thematic relations and are assigned to arguments, while thematic relations are purely semantic relationships; the theta criterion enforces a one-to-one matching between theta roles and arguments, ensuring every argument receives exactly one theta role and every theta role is assigned to exactly one argument, thereby preventing overgeneration in X-bar theory through lexical theta grids that specify a verb's required arguments and their categories.

This lecture introduces the third module of a generative linguistics course, focusing on Bare Phrase Structure and the Minimalist Program. Chomsky transitioned from the Principles and Parameters model to the Minimalist Program to achieve simplicity and economy. The Minimalist Program is not a theory but a program that aligns with computational properties of language. In his 1994 research 'Bare Phrase Structure,' Chomsky presented a new model of grammar. The Economy of Derivation principle states that the shortest derivation is preferred, and overlapping derivations should eliminate unnecessary operations. The Economy of Representation principle states that elements should move to check features in them rather than in the element they move to. Movement is only justified when there is a strong feature requiring checking in the surface structure. Kayne's Linear Correspondence Axiom (LCA) is based on binary branching rather than ternary branching. It addresses issues of linear order in syntax by establishing that the hierarchical structure of phrases determines their linear order. The Minimalist Program represents a significant transition from X-Bar Theory to Bare Phrase Structure, replacing X-Bar Theory with the Merge operation. The directionality of phrase structure construction shifts from top-down (X-Bar Theory) to bottom-up (Minimalist Program). In Bare Phrase Structure, heads are treated as lexical units that can have multiple specifiers, and the distinction between heads and phrases is eliminated.

The minimalist approach to syntax proposes that syntactic structure should be generated without pre-fabricated levels like X-bar, instead using a 'bare phrase structure' where the head element projects directly to form phrases, allowing for multiple specifiers and eliminating the assumption of fixed intermediate levels. This approach challenges the traditional X-bar theory by questioning whether the three-level structure (specifier, intermediate, complement) is necessary or if a simpler structure can account for the same phenomena.

The minimalist program introduced 'merge' as the fundamental operation replacing earlier rule-based approaches, allowing two elements to combine recursively to build complex structures. X-bar theory proposes that all syntactic projections are binary and endocentric, meaning any category X can only combine with one other element, and one of the two must be the head. This restriction greatly improves our ability to explain how children learn syntax efficiently by limiting possible rule formats. The theory applies to all categories (nouns, verbs, adjectives, prepositions) and provides a unified framework. When X combines with its complement (the sibling of the head), it forms an X' projection. Modifiers are siblings to X', differing from complements in that they are adjuncts without direct thematic relationships. This binary structure naturally explains why certain modification patterns are preferred and why others are ungrammatical.

X-Bar Theory is a fundamental framework in generative linguistics that explains how sentences are derived through hierarchical phrase structure, where each grammatical category (X) projects through three levels (XP, X', X) and establishes local syntactic relations with its complement (YP) and specifier (ZP), forming the basis for analyzing sentence structure and drawing tree diagrams in the Minimalist Program approach to language study.

X-bar syntax (1976) revolutionized syntactic theory by introducing intermediate categories (XP, X') between phrasal and lexical levels, enabling more precise hierarchical analysis. The 1986 system built upon this foundation by implementing strict binary branching and replacing the top-level S node with CP (complementizer phrase). Key innovations included P' nodes for prepositional phrases, C' nodes equivalent to former S, and I-bar structure accounting for tense and aspect morphology. The verb phrase now contains only V' and an optional specifier, while determiner phrases (DP) replaced noun phrases to accommodate complex determiners like 'a lot of.'