How Syntax Becomes a Cognitive Engine (Chomskyan Perspective)
We often imagine sentences as something we construct in real time, word by word, like beads on a string.
But in generative syntax, this intuition collapses almost immediately.
A sentence is not assembled.
It is generated.
And what generates it is not vocabulary, but structure-preserving rules operating beneath awareness.
The Central Idea: Syntax Is a Rule System, Not a Word List
Up to this point, we have seen three foundational claims:
Words belong to syntactic categories (N, V, Adj, Det, etc.)Categories define what can combine with what
Meaningful sentences are those that satisfy structural constraints
Now we move one level deeper:
Syntax is not a collection of patterns.
It is a system of formal rules that generate those patterns.
This is where linguistics stops describing language, and starts modeling it.
What Is a Syntactic Rule?
A syntactic rule (or phrase structure rule) specifies:
what a category can consist ofin what order elements combine
and how larger structures are built from smaller ones
For example:
NP → Det + N
This is not a description of language.
It is a generative instruction.
It says:
A noun phrase is built by combining a determiner with a noun.
So instead of memorizing phrases, the mind stores rule systems that construct them.
From Categories to Structure: The Hidden Skeleton of Sentences
Earlier, we tested syntactic categories using substitution:
cat ↔ mouse ✔
the ↔ blue ✘
Now we reinterpret that result structurally.
Take:
“the cat”
This is not just two words.
It is a hierarchical object:
Det → theN → cat
NP → Det + N
So a sentence is not linear.
It is layered architecture.
The First Leap: Phrase Structure Trees
Once rules exist, structure becomes visible.
Instead of:
the cat
We now represent:
NP
Det (the)
N (cat)
This is not notation.
It is a claim about mental representation:
The mind organizes sentences as tree-like structures, not sequences.
Recursion: The Engine of Infinite Language
The most important property of syntactic rules is this:
They can apply to their own output.
Example:
N → Adj + N
So:
cat
furry cat
small furry cat
extremely small furry cat
There is no upper limit.
This is not stylistic creativity.
It is formal recursion.
And recursion is what transforms a finite system into an infinite generative engine.
Co-Occurrence Constraints: Why Words Depend on Each Other
Syntax is not just about order.
It is also about dependency conditions.
Example:
“Sally hasn’t read the book” ✔
“Sally hasn’t read the” ✘
Why?
Because:
Determiners require a following noun.
This reveals a deeper principle:
Words are not independent units.
They are nodes in a constraint network.
Another example:
“three books” ✔
“three book” ✘
Here number imposes morphological agreement.
So syntax is simultaneously:
structural
morphological
and dependency-driven
Word Order Constraints: Structure Over Meaning
Consider:
“Sally saw the book” ✔
“Saw Sally the book” ✘
Nothing about meaning causes failure here.
The failure is structural.
English enforces:
Subject → Verb → Object
So syntax is not meaning-preserving flexibility.
It is rule-constrained positioning of meaning units.
Phrase Structure Rules: The Grammar as a Generative System
We now move from observations to a formal system.
A grammar is a set of rules like:
S → NP + VPNP → Det + N
VP → V + NP
N → Adj + N (recursive)
This system does something profound:
It generates all and only grammatical sentences of a language.
This is the generative hypothesis in its purest form.
Language is not stored.
Language is produced by a rule system in real time.
Lexicon: Where Words Enter the System
Rules alone are not enough.
We also need lexical entries:
N → cat | dog | mouseV → saw | liked | kicked
Adj → small | furry | red
The lexicon is not grammar itself.
It is:
The interface between conceptual meaning and structural generation.
Words are not central.
They are inserted into structure.
Recursive Noun Phrases: Structure Inside Structure
Take:
“the furry cat”
This is:
NP
Det (the)
N
Adj (furry)
N (cat)
Now extend:
“the small furry cat”
Here, recursion builds internal depth.
Meaning is not added linearly.
It is embedded structurally.
Verb Phrase Architecture: The Hidden Complexity of Action
Verbs are not uniform.
They are structurally classified:
1. Intransitive Verbs
2. Transitive Verbs
3. Ditransitive Verbs
4. Sentential Verbs
This reveals a key insight:
Verbs encode structural expectations about the world.
They are not just actions.
They are syntactic blueprints.
Sentence Structure: The Universal Template
At the highest level:
S → NP + VP
This is the backbone of English syntax.
Example:
Sally saw the cat
Breakdown:
NP → SallyVP → saw the cat
NP → the cat
This reveals something deep:
Every sentence is a structured interaction between an entity and a predicate.
Prepositional Phrases: The Expansion Layer of Meaning
Consider:
“the cat in the house”
Structure:
NP → N + PPPP → P + NP
Now recursion expands again:
“the cat in the house on the hill near the river…”
Each PP adds relational structure:
locationdirection
attribution
So syntax is not just about sentences.
It is about relational modeling of reality.
Attachment: Where Meaning Becomes Flexible
Prepositional phrases can attach to:
Noun phrases → “the cat in the house”Verb phrases → “slept in the house”
This creates ambiguity in structure, not meaning.
Example:
“I saw the man with a telescope”
Two structures possible:
man has telescopeseeing done with telescope
This is not ambiguity of words.
It is ambiguity of structure.
The Deep Claim: Language Is a Rule-Governed Generator
At this stage, generative grammar proposes a radical shift:
A language is not a list of sentences.
It is a finite system of rules generating an infinite set.
This leads to three consequences:
1. Creativity is structural
We do not invent sentences; we instantiate rules.
2. Understanding is compositional
We interpret structure, not memorized strings.
3. Grammar is mental
Rules exist in cognition, not textbooks.
Conclusion: The Sentence as a Cognitive Artifact
What appears on the surface as speech is actually:
a recursive derivationa hierarchical structure
a rule-generated object
a real-time computation of meaning
We are not speaking sentences.
We are executing a mental generative system.
Final Question
If every sentence is the output of an invisible rule system…
then what we call “language use” is not expression at all.
Is it possible that speech is simply the surface trace of a deeper syntactic computation happening in the mind?

