ENG509 — Final Term Summary (Lectures 23–39)

📘 Lecture 23 — Further Phrases

📖 Overview: This lecture introduces the foundational concepts of Universal Grammar (UG) from a Chomskyan cognitive perspective, distinguishing between competence and performance. It then explores the hierarchical structure of sentences, using constituency tests, labelled bracketing, and tree diagrams to analyze syntactic units from morphemes to discourse.

🗂️ Topics Covered

The lecture covers Chomsky's cognitive approach to grammar, defining tacit grammatical competence versus performance. It then explains Universal Grammar, its criteria of adequacy (universality, explanatory adequacy, constraint, learnability), and the innateness hypothesis via the Language Faculty. The discussion moves to principles and parameters of UG, including the null-subject parameter, the Wh-Parameter, and the head-first/head-last parameter. Finally, it introduces sentence structure, constituency tests, and methods of syntactic representation like labelled bracketing and tree diagrams, concluding with an introduction to discourse grammar.

📝 Lecture Summary

Topic 133: Introduction to Universal grammar

Chomsky takes a cognitive approach to the study of grammar. The goal of the linguist is to determine what native speakers know about their language (which enables them to speak and understand it) and how this linguistic knowledge might be represented in the mind/brain. In studying language, linguists are studying a specific kind of cognition (i.e. human knowledge). Native speakers know, for example, that the negative counterpart of I like syntax is I don't like syntax, and not I no like syntax; this shows they know how to combine words together to form expressions. Native speakers also know how to interpret expressions, such as the ambiguous sentence She loves me more than you, which has two interpretations. This grammatical knowledge is tacit (i.e. subconscious) rather than explicit (i.e. conscious). A native speaker of English might not be able to answer a question such as 'How do you form negative sentences in English?' consciously, yet they can perform the task. Chomsky terms this grammatical competence: the tacit knowledge of the grammar of their language – i.e. of how to form and interpret words, phrases and sentences in the language. This is contrasted with performance: what people actually say or understand by what someone else says on a given occasion.

🔑 Competence: 'the speaker-hearer's knowledge of his language'. A grammar of a language tells you what is needed to know in order to have native-like competence in the language (i.e. to be able to speak the language like a fluent native speaker). Grammar is concerned with competence rather than performance.

🔑 Performance: 'the actual use of language in concrete situations'. Very often, it is an imperfect reflection of competence. Misproductions and misinterpretations are performance errors, attributable to a variety of performance factors like tiredness, boredom, drunkenness, drugs, external distractions and so forth. Performance is more properly studied within the discipline of psycholinguistics, which studies the psychological processes underlying speech production and comprehension.

The study of the grammatical competence of a native speaker is the study of a cognitive system internalized within the brain/mind which is the product of a 'cognitive organ'. In Chomsky's terminology, the ultimate goal in studying competence is to characterize the nature of the internalized linguistic system (or I-language). A grammar of a language is 'a theory of the I-language under investigation'. Chomsky's ultimate goal is to devise a theory of Universal Grammar (UG): 'The theory of human I-languages ... that identifies the I-languages that are humanly accessible under normal conditions'. In other words, UG is a theory about the nature of possible grammars of human languages.

Topic 134: Universal Grammar

A theory of UG must satisfy several criteria of adequacy. The first is universality: it must provide the tools needed to provide a descriptively adequate grammar for any and every human I-language. The second is explanatory adequacy: a theory should explain why grammars of human I-languages have the properties they do and differentiate them from artificial languages (like mathematics) or animal communication systems (like the tail-wagging dance of bees). A third condition is that the theory be maximally constrained: it should provide technical devices so limited in their expressive power that they can only describe natural languages, not other communication systems. A constrained theory should provide a principled explanation for why certain types of syntactic structure and operation aren't found in natural languages, by supposing that grammatical operations obey certain linguistic principles. A related requirement is that grammars should be as simple as possible, using the minimal theoretical apparatus required. Chomsky refers to the interface with thought systems as the 'conceptual-intentional interface (CI)', and to the interface with speech systems as the 'sensory-motor interface (SM)'. A fourth condition is learnability: it must provide grammars that are learnable by young children in a short period of time. The simpler grammars are, the simpler it is for children to acquire them.

Topic 135: Language Faculty

An acquisition theory is concerned with how children acquire grammars of their native languages. Children generally produce their first recognizable word by around twelve months. For the next six months, there is little evidence of grammatical development. During the single-word stage, their productive vocabulary increases, but there is no evidence of the acquisition of grammar, inflections, or productive word combination. At around eighteen months, children start to make productive use of inflections (e.g., doggy, doggies), inflected verb forms (going/gone), and three-word utterances (Want Teddy). By around thirty months, they can produce adult-like sentences. The central phenomenon is the sudden spurt in multiword speech around 18 months after a long period with no obvious grammatical development. This uniformity and rapidity, Chomsky maintains, is determined by a biologically endowed innate Faculty of Language (FL) or language acquisition program in the brain. This program provides children with a genetically transmitted algorithm (i.e., set of procedures) for developing a grammar based on their linguistic experience (listening to people around them). Language acquisition is an ability independent of general intelligence. The uniformity in grammars developed by different speakers of the same language suggests genetic guidance. The child's linguistic experience is often degenerate, based on adult performance that may be a poor reflection of their competence. A further argument for the innateness hypothesis is that language acquisition is an entirely subconscious and involuntary activity. A critical period for the acquisition of syntax exists before puberty, after which native-like syntactic competence is rarely achieved. Development involves three factors: (1) genetic endowment, (2) external data (experience), and (3) principles not specific to FL (e.g., principles of efficient computation).

Topic 136: Principles of Universal Grammar

Since children can acquire any natural language as their native language, the language faculty must incorporate a theory of UG that enables the child to develop a grammar of any natural language based on suitable linguistic experience. Experience of a particular language L serves as input to the child's language faculty, which incorporates a theory of UG. This implies that certain aspects of child and adult competence are known without experience and must be part of our genetic endowment. These aspects would not have to be learned and will also be universal. Thus, seeking to determine the nature of the language faculty is looking for UG principles. For example, in English question-formation, a non-echo question like Who had he said would do what? involves two movement operations not found in the echo question He had said who would do what? While this is grammatical, preposing the wrong wh-word or auxiliary (e.g., Who would he had said do what?) leads to ungrammaticality. A theory of grammar that posits that grammatical operations are constrained by innate principles of UG minimizes the burden of grammatical learning on the child. It provides a straightforward account of the rapidity of grammatical development and predicts that children will not make certain types of errors.

Topic 137: Parameters

The language faculty incorporates a set of universal principles that guide the child in acquiring a grammar. However, not all aspects of grammar are universal. Some language-particular aspects of grammar, called parametrised aspects of grammar, must be learned. Grammatical learning is limited to these parameters that vary from one language to another. For example, consider the contrast between English and Italian:

• (a) Maria thinks that (they) speak French (the subject is obligatory)
• (b) Maria pensa che parlano francese ('Maria thinks that speak French') In Italian, the verb parlano (speak) has no overt subject, so it must have a null subject (a silent pronoun designated as pro). Italian is a null-subject language, whereas English is a non-null-subject language. Another parameter relates to word order, such as the Wh-Parameter, which determines whether wh-expressions are fronted. In English, wh-words are moved to the front (What do you think he will say?), but in Chinese, they remain in situ (Ni xiang ta hui shuo shenme – 'You think he will say what?'). Another parameter concerns the relative position of heads and complements within phrases. English is a head-first language (e.g., students of philosophy, where students is the head and of philosophy the complement), while Korean is a head-last language (e.g., Muneul dadara – literally 'door close'). The language faculty allows only a binary set of possibilities: a given structure is either consistently head-first or consistently head-last.

📌 Example: Comparing head-first vs. head-last structures:

• English: Close the door (Verb-Object)
• Korean: Muneul dadara (Object-Verb, lit. 'door close')
• English: desire for change (Noun-Complement)
• Korean: byunhwa-edaehan galmang (Complement-Noun)

Topic 138: Sentence Presenting

Syntax, or syntactic analysis, may be defined as: (a) determining the relevant component parts of a sentence, and (b) describing these parts grammatically. The component parts of a sentence are called constituents. This involves breaking down the sentence into its constituents and assigning a grammatical label to each, stating what type of constituent it is. This definition implies a 'top to bottom' analysis, starting from the sentence (the largest unit) and proceeding to clauses, phrases, words, and morphemes (the smallest meaningful unit). The reverse 'bottom to top' analysis shows how constituents combine to form larger units. The hierarchical scale of constituents is:

Sentence ⟷ Clause ⟷ Phrase ⟷ Word ⟷ Morpheme

The double-pointed arrows indicate it may be read 'from left to right' (a sentence may consist of one or more clauses) or 'from right to left' (one or more morphemes may constitute a word).

📌 Example: The hierarchical structure of sentence (1) The snake killed the rat and swallowed it:

• Sentence level: The snake killed the rat and swallowed it
• Clause level: [The snake killed the rat] and [swallowed it] (two coordinate clauses; the second is a reduced form of it swallowed it)
• Phrase level: [The snake] [killed [the rat]] and [swallowed [it]] (one noun phrase, two verb phrases)
• Word level: The, snake, rat, and, it are one-morpheme words; killed and swallowed are two-morpheme words (kill + -ed, swallow + -ed).

Topic 139: Structure

The syntactic structure of a sentence may be represented provisionally by marking off each constituent from sentence level to word level with square brackets: [ ].

📌 Example: Bracket analysis of sentence (1):

• Sentence: [The snake killed the rat and swallowed it]
• Clauses: [ [The snake killed the rat] and [swallowed it] ]
• Phrases: [ [ [The snake] [killed [the rat] ] ] and[ [swallowed [it] ] ] ]
• Words: [ [ [ [The][snake] ] [ [killed] [ [the] [rat] ] ] ] [and] [ [ [swallowed] [ [it] ] ] ] ]

Word and phrase boundaries may coincide, as in John laughed, where John is both a phrase (NP) and a word (noun). Sentence and clause boundaries can also coincide, as in the one-word command Run!, which is structurally one sentence, one clause, one phrase, and one word.

Topic 140: Labeled BracketING

The simple bracketing system is improved by adding an appropriate grammatical label to each pair of square brackets. This convention is called labelled bracketing.

📌 Example: Labelled bracketing for The snake killed the rat and swallowed it:

• [S1 [S2 [NP The snake] [VP killed [NP the rat]]] and [S3 [NP [it]] [VP swallowed [NP it]]] ]
• A more detailed analysis to word level: [S1[S2[NP [DET The] [N snake]] [VP [V killed] [NP [DET the] [N rat]]]] [and] [S3[NP [N it]] [VP [V swallowed] [NP [N it]]]]]

Topic 141: Tree Diagram & PS Rules

A tree diagram is a notational device entirely equivalent to labelled bracketing. It provides the same information about the syntactic structure of a sentence. In a tree diagram, we use terminology like node and branch. An element A is said to dominate B if B is lower down in the tree. Immediate dominance means that there is no other node in between. The lexical items (words) are attached to terminal nodes (the bottom nodes), while all other nodes are non-terminal. A set of instructions called phrase structure rules (PS rules) can sum up the system of expanding one unit into others.

📐 Formula: Phrase Structure Rules (PS rules):

• (17a) S → S-coord-S
• (17b) S → NP-VP
• (17c) VP → V-NP
• (17d) NP → { DET-N / N } (The curly brackets {} indicate a choice between the two options.)

📌 Example: A tree diagram for The snake killed the rat would show S branching into NP (The snake) and VP (killed the rat). The VP would branch into V (killed) and NP (the rat). The NP the rat would branch into DET (the) and N (rat).

Topic 142: Subordination in PS rules

The PS rule (17a) S > S-coord-S is not applicable to all sentences, especially those with subordination where a clause is embedded inside another. In a sentence like I know that the snake killed the rat, the VP must be rewritten as V-S instead of V-NP. The PS rule (17c) must be adapted: VP → V- { NP / S }. In this structure, S2 is a subordinate clause (embedded inside S1), and it is immediately dominated by VP. Ss may also appear inside NPs, as in The snake which killed the rat belongs to our neighbours. This requires adapting rule (17d): NP → { DET-N-(S) / N }, where the round brackets () indicate that S is optional.

Topic 143: SENTENCE AND DISCOURSE

While syntax treats the sentence as the largest unit of description, sentences do not normally occur in isolation. They form part of a larger text (or discourse), which is also organized in a particular way. The sentences of a text follow each other in a logical order, and there are often elements (like pronouns it, he, she or connectors like then) that mark relationships with the context. The rules conditioning the organization of texts are the rules of text grammar or discourse grammar, which are different from the rules of sentence grammar (syntax). This course focuses on the formal structural properties of sentences.

📌 Example: In the text, Soon after breakfast Mary Ann brought in The Times. Mr. Carey shared it with two neighbours..., the word it relates to The Times in the previous sentence.

Topic 144: Constituency tests

Sentences are hierarchically organized into constituents (sentences into clauses, clauses into phrases, etc.). Constituency tests are processes that affect complete constituents, helping to discover what the constituents of a sentence are. For example, in the sentence The snake killed the rat and (it) swallowed it, the goes with snake rather than killed, and it is more closely associated with swallowed. The words are not just a random string but are strictly organized internally with an underlying pattern.

⭐ Key Takeaways

Chomsky's cognitive approach defines grammar as a study of the native speaker's tacit, subconscious competence (knowledge of how to form/interpret expressions) as opposed to performance (actual use). The ultimate goal is a theory of Universal Grammar (UG)—a set of innate, universal principles and binary parameters that describe the grammars of all possible human languages. Key parameters include the null-subject parameter, the Wh-Parameter, and the head-directionality parameter. Syntax involves identifying and hierarchically organizing constituents (from sentences down to morphemes) and representing their structure using tools like labelled bracketing and tree diagrams, guided by Phrase Structure Rules. The innateness of the Language Faculty explains the rapid, uniform, and error-resistant nature of first language acquisition, even from degenerate input.

🧠 Quick Revision Questions

1. What is the difference between competence and performance according to Chomsky, and why is grammar concerned mainly with competence?
2. Explain the four criteria of adequacy (universality, explanatory adequacy, constraint, and learnability) that a theory of Universal Grammar must satisfy.
3. What is the null-subject parameter? Provide an example comparing English and Italian to illustrate this parametric difference.
4. How does a tree diagram differ from labelled bracketing, and what do the symbols {} and () represent in a Phrase Structure Rule (e.g., NP → { DET-N-(S) / N })?
5. Using the sentence I know that the snake killed the rat, demonstrate how a simple Phrase Structure Rule like VP → V-NP must be adapted to handle subordination.

📘 Lecture 33 — Wh-movement II

📖 Overview: This lecture explores the structure and formation of indirect questions (both yes-no and wh-), long-distance wh-movement, and the syntax of relative clauses. It explains how complementizers like whether and if signal indirect yes-no questions, how wh-movement operates across multiple clauses, and how relative clauses function as postmodifiers in complex noun phrases. Understanding these patterns is essential for analyzing complex sentence structures in English syntax.

🗂️ Topics Covered

This lecture covers indirect yes-no questions introduced by whether or if, contrasting them with direct yes-no questions that undergo Subject-Auxiliary Inversion. It then examines indirect wh-questions, showing how wh-movement occurs without SAI in embedded clauses. Long wh-movement is introduced, demonstrating how a wh-element can move across multiple clause boundaries. The lecture then shifts to relative clauses, explaining their structure as postmodifying clauses introduced by wh-words such as who, which, whom, and discussing the choice between stranding prepositions and pied-piping. Finally, it distinguishes between restrictive and non-restrictive relative clauses.

📝 Lecture Summary

Topic 192: Indirect Yes-No Questions

Questions may be direct or indirect. For example, the sentence "One evening he asked her whether he might go home with her" contains an indirect question. In this structure, the phrase one evening is an Adjunct that has been fronted from its normal position after the main clause. The complementizer whether (like that or if) occupies the COMP slot in S̄. Unlike that, the complementizers whether and if indicate that the subordinate clause is a question, and they cannot be deleted.

🔑 Definition — Indirect Yes-No Question: A subordinate clause introduced by whether or if that corresponds to a direct yes-no question but does not undergo Subject-Auxiliary Inversion.

📌 Example: Direct question: "May I go home with you?" → Indirect question: "He asked her whether he might go home with her." In the indirect version, the normal Su-Aux word order (he might) is preserved, and the complementizer signals the question.

Topic 193: Indirect Wh-questions

The sentence "Do you know what happens when one of your tail lights fails?" contains an indirect wh-question, corresponding to the direct question "What happens when one of your tail lights fails?" To form indirect wh-questions, we start from a basic pattern where the wh-phrase is inside the subordinate clause as an Object. Wh-movement then moves the wh-phrase out of S into the COMP position of the subordinate clause.

🔑 Definition — Wh-movement without SAI: In indirect wh-questions, wh-movement occurs, but it is not followed by Subject-Auxiliary Inversion. The auxiliaries remain in their normal position after the Subject.

📌 Example: Underlying structure: Mary wondered [S̄ [COMP] [S John was doing [NP what]]] → After wh-movement: Mary wondered [S̄ [COMP what] [S John was doing ---]]. No SAI occurs here.

💡 Why this matters: Wh-movement and SAI are independent operations. SAI occurs alone in yes-no questions and with fronted negative adjuncts. Wh-movement occurs alone in indirect questions and exclamations (e.g., "What fools they have been!"). They are not necessarily linked to question formation.

Topic 194: Long Wh-movement

Consider the sentence "Who do they think is spying on them all the time?" The wh-word who originates as the Subject of the embedded clause. Long wh-movement moves who first to the COMP of the embedded clause, and then in a second move to the COMP slot in front of the main clause. The fronting of who triggers SAI (with do-insertion) in the main clause.

🔑 Definition — Long Wh-movement: A process where a wh-element moves over more than one clause boundary, leaving a gap in each intermediate position.

📌 Example: Underlying structure: They think [S̄ [COMP] [S who is spying on them all the time]] → After movement: [COMP Who] do they think [S̄ [COMP ---] [S --- is spying on them all the time]]?

Additional examples of long wh-movement: "Who did John say he met --- on the bus?" "Who did Mary tell you she had seen --- on the train?" "At what time do they think the taxi will reach the airport ---?"

Topic 195: Relative Clauses

Subordinate clauses introduced by a wh-word which are not questions are relative clauses. They function as Postmodifiers within complex NPs. Complex NPs have the structure: Det + AdjP (Spec) + N (Head) + S̄ (Postmodifier). For example, "The handsome policeman who directed us to the station" has the structure: [NP The handsome policeman [S̄ who directed us to the station]].

🔑 Definition — Relative Clause: A subordinate clause that postmodifies a Head noun within an NP, introduced by a relative pronoun (who, whom, which, etc.) that binds a gap inside the clause.

📌 Example: Underlying structure of "the girl whom John talked to": [NP The girl [S̄ [COMP] [S John talked to whom]]] → After wh-movement: [NP The girl [S̄ [COMP whom] [S John talked to ---]]]

In wh-question formation, there is a choice between stranding the preposition (extracting only the NP) and pied-piping (moving the preposition along with the wh-word). For example: "the girl whom John talked to" (stranding) vs. "the girl to whom John talked" (pied-piping). Pied-piping is obligatory when the wh-word is a determiner (e.g., whose) or when the wh-word is embedded in a larger PP (e.g., in whose house). For Subject wh-phrases, wh-movement operates vacuously since there is no obvious gap in surface structure.

Topic 196: The Meaning of Relative Clauses

Postmodifying (relative) clauses may help to identify the referent of the Head noun by narrowing down its reference. Restrictive relative clauses provide information that is essential for identifying the referent. For example, in "the man who lives next door," the clause who lives next door indicates exactly who is being referred to.

🔑 Definition — Restrictive Relative Clause: A relative clause that provides essential information for identifying the referent of the Head noun; without it, the reference would be unclear.

🔑 Definition — Non-restrictive Relative Clause: A relative clause that provides additional, non-essential information about the Head noun; it does not narrow down the reference since the referent is already sufficiently clear.

📌 Example: Restrictive: "The books which I ordered have finally arrived." Non-restrictive: "The Pope, who was expected to visit Poland again this summer, decided to go to France instead."

💡 Why this matters: The distinction between restrictive and non-restrictive clauses also applies to other types of postmodification and premodification, and has implications for comma usage, pronoun deletion, and the use of that.

⭐ Key Takeaways

This lecture establishes several core principles of wh-movement and relative clause formation. First, indirect yes-no questions use complementizers (whether/if) and crucially do not undergo SAI, unlike their direct counterparts. Second, indirect wh-questions involve wh-movement without SAI, proving these two operations are independent. Third, long wh-movement can move a wh-element across multiple clause boundaries, leaving gaps in each intermediate COMP position. Fourth, relative clauses are formed by wh-movement of a relative pronoun into the COMP of a postmodifying clause, with optional choice between stranding prepositions and pied-piping (the latter being obligatory in certain contexts like whose). Finally, relative clauses are classified as restrictive (essential for identification) or non-restrictive (providing additional information), a distinction that affects pronoun deletion and comma usage.

🧠 Quick Revision Questions

1. What is the key difference between direct and indirect yes-no questions in terms of Subject-Auxiliary Inversion?
2. Explain how wh-movement and SAI can occur independently of each other, providing one example for each operation occurring alone.
3. How does long wh-movement differ from regular wh-movement, and what is the role of the intermediate COMP positions?
4. Under what circumstances is pied-piping obligatory in relative clause formation?
5. What is the fundamental difference between restrictive and non-restrictive relative clauses, and how does this distinction affect the use of the complementizer that?

📘 Lecture 39 — Control: Non-lexical subjects

📖 Overview: This lecture examines how non-finite subordinate clauses handle subjects that are not overtly expressed, focusing on control structures. It distinguishes control from raising verbs, explores subject versus object control, and introduces reanalysis as a process of structural ambiguity in verb-particle combinations. Understanding these concepts is essential for analyzing sentence structure in English syntax.

🗂️ Topics Covered

The lecture begins by examining non-lexical subjects in non-finite clauses and the convention for marking empty subject positions. It then distinguishes control verbs from raising verbs using three diagnostic tests. Subject and object control are differentiated based on which NP controls the empty subject. Indefinite control is introduced for patterns with no overt controller. The lecture concludes with control adjectives and an introduction to reanalysis, focusing on the structural ambiguity of verb+preposition strings and the distinction between prepositions and particles.

📝 Lecture Summary

Topic 229: Control: Non-lexical subjects

Sentences like I did not wish to write a family saga contain a non-finite object clause whose subject is not lexically filled. The convention for marking such non-lexically realised subject positions is:

(2) I did not wish [sˉ [s ( ) [VP to write a family saga ] ] ]

Non-finite subordinate clauses may have a lexical subject, as in I did not wish [sˉ(for) [s the book to become a family saga]]. The complementiser for is allowed here but disallowed in (2) because the string for...to is not grammatical in Standard English.

In (1), an NP outside the subordinate clause controls the subject position in the subordinate clause. The non-lexical subject is controlled by the main clause subject NP (I): ( ) is interpreted as 'I'. Reflexivisation supports this analysis.

🔑 Definition — Control: The syntactic relationship where an NP outside a non-finite clause determines the reference of an empty subject position within that clause.

📌 Example: (4) I want [sˉ [s ( ) to enjoy myself ]]. The reflexive myself is bound inside the subordinate clause by the non-lexical subject ( ), which is controlled by the higher subject I.

(6) I want [sˉ you to enjoy {yourself /*myself}]. Here, you is the overt lexical subject, so myself is ungrammatical.

Topic 230: Control vs raising

It is important to realise that the controller NP in (5) is both the underlying and the surface subject of want; I has not been moved into subject position by a raising transformation. Verbs like want, decide, wish, hope are control verbs, not raising verbs.

Raising verbs (seem, appear, etc.) appear in the string: NP-seem --- to-infinitive, where --- cannot be filled by a lexical NP: (7a) They seem --- to have left. (7b) *They seem the children to have left.

Three diagnostic tests distinguish raising from control:

1. Lexical Subject in non-finite clause: Raising = No; Control = Yes
2. It-extraposition paraphrase: Raising = Yes; Control = No
3. There V to...: Raising = Yes; Control = No

📌 Example: For raising verb seem: (9) There seem --- to be many children with reading problems. For control verb want: (10) I want {()/You} to leave. And (12) *There want () to be many children with reading problems.

💡 Why this matters: These three tests provide a reliable method to determine whether a verb is a raising verb or a control verb, which is crucial for understanding the syntactic structure of sentences.

Topic 231: Subject and object control

(13) He promised her to mow the lawn. (14) General Dyer told his troops to fire on an unarmed crowd in the streets.

Both sentences have the pattern Su [V-Oi-Od], unlike sentence (1) which has Su [V-Od]. In (13), the non-lexical subject is controlled by the Subject NP He, not by the NP/Oi her. In (14), it is the NP/Oi his troops which is the controller, not the NP/Su General Dyer.

The type of verb in the main clause determines control type: Verbs like promise have Subject control, whereas verbs like tell, order, ask, force, persuade, allow, encourage have Object control.

(15) [s [NP He][ VP promised [NP her ][sˉ ( ) to mow the lawn ]]] (16) [S [NP General Dyer] [VP told [NP his troops] [sˉ ( ) to fire on an unarmed crowd in the streets] ] ]

Sentences with object control can be passivised: the NP/Oi can be made the NP/Su of the passive sentence.

📌 Example: (21)[s [NP He] [VP persuaded [NP her] [ ( ) to buy some more books ] ] ] → (22)[s[NP She][VP was persuaded ---[s ( ) to buy some more books] ] ]. Now the subject she controls the empty subject position.

Topic 232: Indefinite control

English also has control patterns with no overt controller.

(23) {To write/writing} a book on this subject was an irrational act. (24) [sˉ [s () {to write/writing} a book on this subject] ] was an irrational act.

The controller might be somebody mentioned in context, or someone unknown or indefinite — this is indefinite control.

Other examples: (25) Is ( ) to say No to nuclear weapons a defence policy? (26) Oxbridge's grip on the top ranks of the civil service should be broken by ( ) encouraging students at provincial universities ( ) to come forward as potential mandarins.

Non-finite clauses may function as Adjuncts or Modifiers: (27) I want to go to Paris to buy some new clothes. (The empty subject position is controlled by I) (28) He received a letter ordering him to leave the country. (The empty subject is controlled by a letter)

Topic 233: Control adjectives

There are also control adjectives in English, not just control verbs. These adjectives are followed by non-finite clauses with controlled subject positions:

(29) I am eager [sˉ [s( ) to come to your party] ] (30) He is reluctant [sˉ [s ( ) to tell her the truth]]

Applying the three tests confirms these are control patterns, not raising patterns:

(31) I am eager for Mary to be at the party too. (Lexical subject possible) (32) *It is eager that I should come to your party. (No it-extraposition) (33) *There is eager to be a party next weekend. (No there raising)

Topic 234: Introduction: Reanalysis a problem of bracketing

Reanalysis (also called 'restructuring') changes the structure of a string of elements, so that before and after reanalysis the string will pattern differently.

Examples: bring in, bring out, kick the bucket (idiom).

The idiom kick the bucket can regularly occur in the passive (The bucket was kicked) but not in its idiomatic sense: *The bucket was kicked by the old man (meaning "The old man died"). As an idiom, it has been restructured into a composite intransitive verb.

Strings such as depend on, look at, give in to may be structurally interpreted as intransitive verbs followed by a PP, or alternatively as transitive verbs followed by a Direct Object. These kinds of structural ambiguity result from reanalysis.

⭐ Key Takeaways

The distinction between control and raising verbs is fundamental to understanding non-finite clause structures, with three diagnostic tests (lexical subject possibility, it-extraposition, and there raising) reliably separating them. Control structures involve an NP outside the clause determining the reference of an empty subject, and this controller can be the subject (subject control with verbs like promise) or the object (object control with verbs like tell). Indefinite control occurs when there is no overt controller, and control adjectives follow the same patterns as control verbs. Reanalysis introduces structural ambiguity in verb+preposition strings, where elements may be interpreted as either part of a phrasal verb or as a separate prepositional phrase, requiring careful testing to determine their true constituent structure.

🧠 Quick Revision Questions

1. What are the three diagnostic tests that distinguish control verbs from raising verbs, and how do the results differ for each type?

2. In the sentence He promised her to mow the lawn, which NP controls the empty subject position, and what type of control is this?

3. What is indefinite control, and in what kind of sentence structure does it typically occur?

4. How does the passivisation of an object control sentence affect the controller of the empty subject position?

5. What is reanalysis in syntax, and how does it create structural ambiguity in verb+preposition strings?