BASIC ISSUES IN SYNTAX

EFL University
2023

P Madhavan
The Structure of Language
• There is much in common among languages of the world. And there
are differences, too.
• Syntax studies the ways in which words of a language are put together
to form sentences.
• Words combine to form phrases, phrases combine to make clauses,
and clauses combine to give sentences.
• A sentence is the upper limit for syntactic analysis.
• Anything larger than a sentence is a discourse, which falls outside the
scope of syntax.
What is a word?
• A word is not easy to define. So we will simply assume that we know
what is a word.
• Some tests:
• Are eat, eats, ate, eating and eaten different words or different forms
of the same word?
• Is “bookshop” one word or two words?
• If two words A and B mean the same , are they two words or one?
• gagane in Sanskrit means ‘in (the) sky’, but it is treated as single word.
• ‘brother-in-law’: one word or three words? What is the plural?
Phrase
• A phrase is between the level of words & the level of
sentences/clauses.
• e.g., [[I] [love [linguistics]]] is a clause, consisting of three words. How
many phrases? Three.
• The subject “I” is a word as well as a phrase.
• “love linguistics” is a phrase (consisting of two words).
• “linguistics” is a single-word phrase.
More on Phrases
• Typically, a phrase consists of more than one word, some examples:
• [books [on linguistics]], [a chapter from a novel] – both these are
noun phrases, NPs. [on linguistics] is a PP, a prepositional phrase.
• It is possible to have a phrase within a phrase, as in the above
examples.
• When one phrase is inside another, it is called “embedding” or
“recursion”.
• Recursion is a distinctive property of human languages/natural
languages.
Distribution
• Why do we treat [books [on linguistics]] as a phrase?
• The structuralist linguistics stressed the notion of “distribution”.
• The question to ask is ‘where does the phrase occur?’. Consider the
following sentences:
• i) Books on linguistics are, generally, boring. (subject position)
• ii) I do not like books on linguistics. (object position)
• The distribution of this phrase is the set of positions where an NP can
occur.
• Let us take an example from English phonology to get an idea of what is
meant by “distribution”.
Distribution
Consider the sounds /k/, /t/ and /p/ in English. All the three can occur
at the beginning, middle and end of a word.

What about the sounds /ŋ/ and /h/? There’s no word that begins with a
/ŋ/. Is there a word that ends in /h/?

The set of environments where a sound can occur its distribution. And
you can generalise this statement to any linguistic item.
Phrases
• A phrase must have one and only one “head”.
• The head is usually a grammatical category like noun, verb,
preposition, adjective or adverb.
• A phrase is known by the head, thus we have NP, VP, PP, AP (i.e.,
adjective phrase), and Adv P.
• Also the head must be inside the phrase, this is known as the
property of “endocentricity”.
• A phrase can be diagrammatically represented by means of a “tree
diagram”.
 Immediate constituent analysis
Structuralist linguists employed a tool known as IC analysis to represent the
structure of a phrase.
Take a phrase “ a tale told by an idiot”

a tale told by an idiot

a tale told by an idiot

by an idiot
an idiot
Tree diagram & nodes
• NP → Det Adj N The arrow is read “goes to”
• e.g., an interesting book
NP VP → V NP

Det Adj N VP

an interesting book V
NP

VP → V NP saw a kite
e.g., saw a kite
 Prepositional phrase
• [on linguistics] is a prepositional phrase.
PP PP → P NP

P NP

on linguistics
Postpositional phrases have a structure which will be a mirror image of this,
PP → NP P
e.g., Malayalam: mesa mel PP complement
‘on table’
NP P

mesa mel
Adjective phrase AP
• AP → A PP e.g., “proud of my daughter”
AP
A PP
P NP
Det N
proud of my daughter an obligatory complement
 more complex NPs
e.g., a book on introductory linguistics in English
[NP [Det[a] ] [N’ [N book] [PP on [NP[A introductory ] [Nlinguistics ]]] [PP in [NP English]]]]

Colour coding: [ brackets for the whole NP, [ for N’, [ for PP, and [ for terminal items.
The number of opening brackets must match the number of closing brackets, 10 each
here.
PS Rules Phrase Structure Rules
NP → Det N’ N-bar
N’ → N PP PP
PP → P NP
Tree diagram - nodes and association lines
NP
Det N’

N PP PP

P NP P NP

A N
N

a book on introductory linguistics in English

Ā or A-bar
Linear and hierarchical relations
• A tree diagram encodes two kinds of relations between words in a
phrase/sentence, the linear (precedence) relation and the hierarchical
relation.
• This works as in a family tree – a root, and branches and leaves at the
end of the branches.
• Nodes and association lines
• Preterminal strings and terminal strings.
Disambiguation using tree diagrams
1. Old men and women are eligible for pension.
a) [[Old men and women] are eligible for pension.]
b) [[[Old men] and [women]] are eligible for pension.]

2. I watched the policeman with a telescope/standing on the terrace.

a) [I [VP watched [NP the policeman] [PP with a telescope]]]
b) [I [VP watched [NP[NP the policeman] [PP with a telescope]]]]
Head and complement
• Apart from the head, a phrase may have complements and adjuncts (or
modifiers). These can be phrases, too. A transitive verb obligatorily takes
an NP as its complement, a preposition also takes an NP complement.
• A noun may also have a complement, e.g., destruction of the city by the
enemy. In this, ‘destruction’ is the head N, ‘of the city’ is the
complement and ‘by the enemy’ is an adjunct.
The enemy destroyed the city.
He criticized the book – his criticism of the book.
• Some adjectives like ‘proud’ are transitive, they require a complement.
e.g., I am proud [of my daughter].
 Order of constituents
• a student of physics with long hair of physics = complement
• *a student with long hair of physics with long hair = adjunct
NP NP*
det N’ PP det N’
PP

N PP N PP

P NP P NP P NP P
NP
N N
a student of physics with long hair a student with long hair of physics
Noun and noun phrase
• The noun in an NP is its head, it may or may not have complements
and modifiers.
• This in turn implies a lone noun when it appears in a syntactic string is
to be treated as an NP.
e.g., I do not like linguistics.
• The underlined word is an NP, not an N, syntactically speaking.
• Similarly, a single verb can be a VP.
e.g., They laughed.
If the verb is transitive, the VP must contain the object NP.
The English Determiner
• As you know, English is a language with an article system. The definite
article has only one form the (though in pronunciation, there are two
variants /δə/ & /δi/ ). The indefinite article has two forms: a and an.
• *my this book / *this my book, *this the book / *the this book, *the
my book / *my the book are all ungrammatical.
• There is only one slot in Det, which can be occupied by either an
article, or a possessive or a demonstrative. [ this, these, that, and
those ]
William Jones - the discovery of Sanskrit
The Sanscrit language, whatever be its antiquity, is of a wonderful
structure; more perfect than the Greek, more copious than the Latin,
and more exquisitely refined than either, yet bearing to both of them a
stronger affinity, both in the roots of verbs and in the forms of grammar,
than could possibly have been produced by accident; so strong indeed
that no philologer could examine them all three, without believing them
to have sprung from some common source, which perhaps no longer
exists: there is a similar reason, though not quite so forcible, for
supposing that both the Gothick and the Celtick, though blended with a
very different idiom, had the same origin with the Sanscrit; and the old
Persian might be added to the same family.
Comparative Philology
Jones summarized his findings from a six-month long study of
Sanskrit on February 2nd, 1786. His findings launched not just
academic careers, but also furthered the imaginations of
political, social and chauvinistic programs for more than two
centuries. The debate continues to this day with geneticists and
cultural recidivists making claims that seek to upend or confirm
the implications of Jones’ original and striking claim that:
Langue and parole
Formal criteria in defining word classes
The traditional grammars used meaning criteria to define word classes
like the noun and the verb.
Structuralists relied on formal criteria.
For example,
a noun can be preceded by a determiner.
a noun can take a plural inflection.
a verb can take a tense affix.
an adjective can take a comparative and a superlative affix.
a preposition is indeclinable.
The English auxiliary system
• Look at the tense in English.
John writes/ is writing/ has written/ has been writing a letter to his
mother.
John wrote/ was writing/ had written/ had been writing a letter to his
mother.
John will write/ will be writing/ will have written/ will have been writing
a letter to his mother.
What is the most economic way of describing this set of sentences?
Aux expansion
• Aux Tense Modal Prog(ressive) Perf(ective)
• Pred P Aux VP
• VP V NP
Tense: Pres(ent) / Past
Modal: can, may, must, will, shall, ought to, have to etc. and do
Prog: be-ing
Perf: have-en (-en stands for the past participle ending, which may be
variously realized as –ed as in ‘walked’, en as in ‘written’, zero as in ‘put’
by vowel change as in ‘sung’, and so on)
The rule of Affix-Hop
Using a single rule of Affix-hop, all these forms can be derived.
Affix X X + Affix , where Affix is any item under the Aux node that
is affixal in nature (i.e., Tense, -en and –ing), and X is any non-affix.
writes = Tense (Pres) write = write + pres =‘writes’
is writing = Tense be-ing write = be + pres (is) & write +ing (writing)
has written = Tense have-en write = have + pres (has) write + en
(written)
has been writing = Tense have-en be-ing write =
have + pres (has) be + en (been) write + ing (writing)
The rule can generate all verbal forms
• T will have been writing
pres will have-en be-ing write Affix X = X + Affix
After Affix Hop, we get
will + pres (will) have be+en (been) write+ing (writing)
Notice that ‘have’ remains as ‘have’ because no Affix has attached to it.
The Affix-Hop rule applies (takes effect) more than once, in fact, as
many times as it can, over a given string. The Affix moves over the next
item and attaches to its right end, as shown in the diagram.
Identifying a clause
• In traditional grammar, a clause is identified with the occurrence of a
finite verb. A finite verb is one with tense and agreement.
• For example, ‘She left the teaching job and joined an IT company’.
• The above sentence has two finite verbs (underlined), hence has two
clauses.
Afterleaving the teaching job, she joined an IT company.
• The sentence above has a nonfinite verb. Hence how do we treat
‘after leaving the job’, as a clause or a non-clause?
• The answer is, YES, it’s a clause.
What is a clause?
A clause consists of a subject and a predicate.
This definition you can trace back to Aristotle.
Both these terms - subject & predicate - are not easy.
We will come back to these terms, later.
 a verb, whether finite or not, projects a
clause
• The subject may be overt or covert.
The English verb has three non-finite forms: the bare verb, the ‘to +
verb’ and the ‘verb + ing’, as shown below.
I watched her cross the road.
She was asked to resign.
I saw John fishing in the lake.
The finite verb forms are many. See the slides on the Affix-Hop.
Embedding of clauses
• When a clause is embedded within another clause, a “connector” is
employed, as shown below:
• John said, “Peter loves Mary”.
• John said that Peter loves Mary.
• In this, ‘that’ is the connector; it is called a complementizer because it
links the complement clause to the matrix clause.
• In the above example, “that” is optional, though not in all cases.
• John discovered/ reported that Peter loves Mary.
A complementizer phrase - CP
• Before we think of the projection called CP, let us see the structure of
the matrix clause. What is the head?
•S NP VP was a PS rule in the early days, but later it was
realized that this is in violation of the principle of endocentricity.
• The head of a clause is inflection. That is, clause = IP.
• IP NP Infl VP
• The Infl is supposed to contain Tense & Agreement.
• CP C IP (In Indian languages, C comes after the IP)
The imperative sentence
• The imperative sentence in English usually lacks an (overt) subject.
There is however ‘an understood subject’, a second person you.
• How do we construct a syntactic proof for this?
1. He washed himself. * Wash himself.
2. She washed herself. * Wash herself.
3. I washed myself. * Wash myself.
4. They washed themselves. * Wash themselves.
5. We washed ourselves. * Wash ourselves.
6. You washed yourself. Wash yourself.
 A reflexive need to have an antecedent in
the clause
• If we take the imperative sentence to have a unique ‘you’ as the
subject, the ungrammaticality of the reflexive sentences of column
two receives a straightforward explanation. It also explains why the
last sentence is found to be grammatical.
•
Another example
• Hold your/*his/*my/*our/*her/*their breath.
• He held his breath/ we held our breath/ she held her breath/ they
held their breath/ I held my breath/ you held your breath.
• One can hold only one’s own breath, not others or not of other’s.
• ‘Hold one’s breath’ is a phrase that behaves like a reflexive.
Constituency Tests
• Only constituents can undergo syntactic operations, not part of a
constituent. There are standard tests for determining whether a given
string is a constituent or not.
A. Fragment answers:
1. This is [my favourite book]. What is it? My favourite book. NP
2. The cops chased the thief. What did the cops do? Chase the thief. VP
3. The book is on the table. Where is the book? On the table PP
4. The teacher put the book on the table. What did the teacher do? ……VP
5. I want some of the students to graduate in Linguistics. Who do I want to
graduate in Linguistics? Ans: *some of
Constituency Tests
B. Coordination – only constituents of the same type can be
coordinated.
1. [Rahul] and [(two of) his friends (from the US)] visited us. NPs
2. She [opened the book] and [read the first page]. VPs
3. The race car went [[over the bridge] and [through the tunnel]]. PPs
4. The fresher looked [extremely nervous] and [out of his depth]. APs
Constituency Tests
• C. Substitution (the part substituted is given in square brackets) - pro-forms
1. [My uncle] is very resourceful, he might be able to help you.
2. [The lady speaking now] is my aunt. She *(speaking now) is my aunt.
3. I found [a book about black magic] and I read it *(about black magic)
4. Rahul [went home early] and so did his sister *(early).
5. The teacher chose [[a student] from Linguistics] and one from English.
6. Roger [opened the box [with a box cutter]] and so did Rahul with a knife.
7. Roger [went [to the post office] and Rahul to the hospital.
[We will get back to the examples 5-7 which employ ellipsis.]
Ellipsis
• It is often possible to elide certain chunks of expression in sentences,
when said in a speech context. This phenomenon is known as “ellipsis”.
• Tag questions is an example of ellipsis.
• You plan to go to the party tomorrow, don’t you [plan to go to the party
tomorrow]?
5. The teacher chose a student from Linguistics and one [ a student] from
English.
6. Roger opened the box with a box cutter and so did [open the box] Rahul
with a knife.
7. Roger went to the post office and Rahul went to the hospital.
Topicalization (i.e., movement)
1. I can bet [that guy in a red shirt] is a lawyer.
That guy in a red shirt, I can bet, is a lawyer.
* That guy in, I can bet, a red shirt is a lawyer.

2. The cricketer gifted [his newly acquired brand new car] [to his
girlfriend].
His newly acquired brand new car, the cricketer gifted to his girlfriend.
To his girlfriend, the cricketer gifted his newly acquired brand new car.
* His newly acquired, the cricketer gifted brand new car to his girlfriend.
Clefting & pseudoclefting
I watched the actor’s funeral with a heavy heart on television.
It is [with a heavy heart] that I watched the actor’s funeral on television.
It was [the actor’s funeral] that I watched with a heavy heart on television.
It is [on television] that I watched the actor’s funeral with a heavy heart.
It is [I ]who watched the actor’s funeral with a heavy heart on television.
*It is I watched that the actor’s funeral with a heavy heart on television.
• The structure of a cleft is
It is [ the clefted item] that ……
What I watched with a heavy heart on television was [the actor’s funeral].
The Lexicon
• Lexicon is the repository of all idiosyncratic information (on words) in
a language.
• Such information includes meaning of the words, their spelling and
pronunciation, also any particular aspect of the word’s use. Whatever
is not systematic and predictable, in other words.
• Take the plural formation. Adding the suffix /–s/ is predictable. But,
the plural form ‘children’ is idiosyncratic. So with irregular past tense.
• In synthetic languages much of the inflectional forms are
unpredictable, or only partly predictable. So they have to be provided
in the lexical entry.
Lexicon is the input to syntax
• Words chosen from the lexicon is the input to syntax. But the inflection on
the words will be added in syntax.
• For example, take the subj-verb agreement in English.
• We/she read/ reads the Bible every day.
• And it is also true that a word gets its meaning in a syntactic context. This is
especially true of polysemous words.
• e.g., We are planning to give it a dry run/ The batsman scored 39 runs/ This
road runs straight for another two miles/ Donald Trump is running for the
Presidency/ This fabric is good, the colour doesn’t run/ He has a bushy
eyebrow, a trait that runs in the family/ In the long run it makes sense, etc.
Idioms
• The meaning of an idiom is unpredictable. For example, “kick the
bucket”, “hang up the shoes”. So they need to be given in the lexicon.
• The words in an idiom need to stay together for the idiomatic
meaning. If you passivize the idiom “kick the bucket” to “the bucket
was/is kicked” the idiomatic meaning is lost, it can have only the
literal meaning.
Where is grammar?
• The word ‘grammar’ is used in different senses.
• We will mean by ‘grammar’ the sum total of organizational principles
that govern the use of a particular language.
• By “the grammar of the English language” we do not mean what is
printed in the book by Quirk and Greenbaum, for instance. The
implicit knowledge of a native speaker regarding the use of his/her
language is what we mean by grammar.
• Were you taught the grammar of your mother tongue?
• Does your mother tongue have a grammar?
Evidence from Language Acquisition
• The mystery of language acquisition in young children
• LAD – Language Acquisition Device
• The “Poverty of stimulus” argument – also called Plato’s problem
“How comes it that human beings, whose contacts with the world are
brief and personal and limited, are nevertheless able to know as much as
they do know?”
The innateness hypothesis: The language faculty is innate in humans. In
other words, humans are ‘pre-programmed’ to learn language. All that is
necessary is adequate exposure.
Language is a mental organ.
The Chomskyan Program
• Noam Chomsky introduced a new approach to the study of grammar,
which came to be known as Transformational Generative Grammar.
• Chomsky contested the then prevalent structuralist practice which
relied heavily on the immediate constituent analysis [IC analysis].
• IC analysis cannot reveal the finer aspects of sentence structure,
aspects that the native speaker is implicitly aware of; a descriptively
adequate grammar ought to, Chomsky argued, represent the innate
knowledge of the native speaker, and ascribe to “all and only”
grammatically well-formed sentences a structural description.
 Easy to please
• He drew attention to pairs of sentences like
i) John is easy to please.
ii) John is eager to please.
IC analysis will give identical diagrams to (i) and (ii). [see the next slide]
But, the interpretation of these sentences differ. In (i) who is pleased is
John, in (ii) John is the one who is pleasing someone.
iii) John promised Bill to leave iv) John persuaded Bill to leave.
In (iii) it is John who leaves, in (iv) it is Bill.
The question is how can this difference in reading be shown in the
structural description/representation.
Analysis trees
S S

NP VP
NP VP

V NP clause V
clause

John persuaded/ Bill [ to leave] John is easy/ [ to please]

promised eager
Deep structure and surface structure
• The difference in interpretation in these sentences can be shown if we
resort to a two-level syntactic analysis: a deep structure and a surface
structure.
a) [It is easy [(for) someone/anyone to please John]] deep structure
b) [John is eager [ John to please someone]] deep structure
John is easy/eager to please surface structure
c) [John promised Bill [John to leave]] deep structure
d) [John persuaded Bill [Bill to leave]] deep structure
John promised/persuaded Bill to leave surface structure
Transformational Rules
• The deep structure, which is closer to meaning, is mapped on to the
surface structure by applying “transformations”.
• Syntactic Structures (1957) gives a sketch of this two-level approach
to the syntactic analysis of English sentences.
• Aspects of the Theory of Syntax (1965) is the book that launched the
Chomskyan program in Linguistics. The theory presented in this book
came to be known as the Standard Theory. Also known as the
“Aspects Model”.
• See the diagram
The Aspects Model 1965
• Flow chart
Lexicon deep structure
meaning
PS rules

T- Rules
phonological
rules

surface structure

phonetic form
Transformational rules
• In the 1965 model, T-rules were employed for mapping deep
structure to surface structure.
• An example of a T-rule is the ‘subject-auxiliary inversion’ in English
interrogatives.
• John can win the race. Question: Can John win the race?
initial string: John Tense can win the race
affix hop: John can+Tense win the race
aux inversion: can+Tense John win the race
Can John win the race?
Do support
• John won the race. Question: Did John win the race?
• John past win the race initial string
past John win the race aux inversion
past do John win the race do support
do+past John win the race affix hop
Did John win the race? Question form
When the aux is just Tense (with no Modal/Prog/Perf), for questions
and emphatics an additional rule that inserts a ‘do’ immediately to the
right of Tense is invoked. This is known as ‘do support’.
Autonomy of syntax
• Chomsky advanced the thesis of “the autonomy of syntax”.
Essentially, what it states is that syntax has its own principles, it’s not
driven by meaning considerations.
• Colorless green ideas sleep furiously.
• This famous Chomsky sentence illustrates this point. The sentence is
well-formed; it obeys the rules of sentence grammar. The oddity
about the sentence is to be addressed by the meaning component of
grammar.
• Notice that a jumbled form of the above sentence is ungrammatical.
• *Green sleeps colorless furiously ideas.
Adequacy : observational /descriptive/
explanatory
• Levels of adequacy are design criteria that have played a central role in the
transformational generative grammar model since its inception in the 1950s.
• According to Chomsky, an adequate model of language not only must state what
combinations of expressions do and do not give rise to grammatical sentences
(observational adequacy) but must also account for the knowledge system
underlying the intuitions of the native speaker (descriptive adequacy) and explain
how such knowledge is acquired (explanatory adequacy).
• Explanatory adequacy chooses the optimal grammar from among competing
descriptively adequate grammars.
• The tension between the goals of descriptive adequacy and explanatory
adequacy has led to important developments in the Chomskyan model over the
past half century.
The Chomsky Hierarchy of Grammars
• Chomsky drew the attention of intellectuals much earlier by
proposing the “Chomsky hierarchy of grammars” which is still
considered one of the basic planks of the modern computer science.
• Type 0 grammar /Type 1 grammar/Type 2 grammar/ Type 3 grammar.
• Type 0 is an unrestricted grammar which outputs a “recursively
enumerable” language. Type 1 is a context-sensitive grammar which
outputs a context-sensitive language. Type 2 is a context-free
grammar which outputs a context-free language. Type 3 is a regular
grammar that outputs a regular language, that is, the class of human
languages/natural languages.
Hierarchy of grammars & Automaton
• Type 0 Turing Machine

• Type 1 Linear bounded

automaton

• Type 2 Push down

automaton

• Type 3 Finite state automaton

Review of “Verbal Behavior”
• In 1959 Chomsky wrote a review of B. F. Skinner’s Verbal Behavior
(1957) which catapulted him to fame. It was a scathing attack on
some of the cherished principles of behaviorist psychology. He argued
that language learning does not follow the operand conditioning, of
the stimulus-response pattern. This review has been a turning point
that inaugurated a series of developments, culminating in the
cognitive revolution in social sciences.
The poverty-of-stimulus argument
The data the child is exposed to is flawed and highly ‘impoverished’.
Speech is often marred by phenomena like sentence fragments,
unfinished utterances, online corrections, memory lapses, repetitions
and so on. Yet the child is able to arrive at the core grammar of the
language, which is nothing short of a miracle. The concept of Universal
Grammar (UG, for short) is thus inevitable, notes Chomsky.
“The ideal native speaker-hearer” is an ancillary theoretical construct
that is invoked to explain language acquisition in infants.
Flying planes can be dangerous
The ambiguity of this sentence is not due to the presence of a lexical
item that has two different senses. It is an instance of structural
ambiguity. In one reading, flying is an adjective, and in the second
reading, flying is a verb with an unspecified subject.
If we replace ‘can be’ with ‘is/are’ the ambiguity disappears.
A similar sentence is ‘Visiting relatives can be a nuisance’.
These examples also argue in favour of a two-level analysis of the
clause, in terms of a deep structure and a surface structure.