Week 2 – Constituency, Distributional Analysis & Syntactic Categories

Overview

  • Week two lecture of Syntax course: focus shifts from individual words to larger structures (phrases/constituents) and to the reasoning tools syntacticians use.
  • Tutorials also begin this week; practical application of concepts is emphasized.
  • Central theme: constituency—how grammar groups words into units that behave together.
  • Methodology throughout the course: distributional analysis—classifying elements by where and how they appear (internally & externally).

Quick Revision – Word Classes & Categories

  • Every lexical item belongs to a category (noun, verb, adjective, adverb, preposition, determiner, etc.).
  • Categories are language-specific; identical labels across languages mask different defining behaviors.
    • E.g.
    • English “noun” ≠ Ngongonbu “noun” in exact feature set.
    • Some languages lack prepositions or determiners entirely.
  • As descriptive linguists we must discover the categories for each language; cannot assume the English inventory.
  • Key analytic tool: distributional analysis (internal vs. external evidence).

Distributional Analysis

Internal (Morphological) Evidence

  • Look inside a word:
    • Inflectional markings (number, tense, aspect, mood, gender, etc.).
    • Presence of known derivational morphemes (suffixes/prefixes) that reliably change category (e.g. -ness → adjective → noun).
    • Note: Internal evidence is language-specific; e.g. English plural \text{-s} appears only on nouns, but some nouns (sheep, ox) lack it, so absence ≠ non-noun.

External (Syntactic) Evidence

  • Examine how the phrase headed by the word behaves in larger structures:
    • Which syntactic functions can it fulfil? (subject, object, predicate, modifier, etc.).
    • What dependents or modifiers can attach to it?
    • Only NP’s or clauses serve as subjects in English → if X is subject, X is an NP or a clause.

Worked Example – “The ferocious possums fled from the scene of the crime”

  • Internal:
    • possums bears plural \text{-s} ⇒ typical of English nouns.
  • External:
    • Heads the subject phrase; only NP/clauses appear here ⇒ likely noun.
    • Dependent words the, ferocious cannot remain when “possums” is removed → they depend on the noun.
  • Substitution test within category:
    • Replace with cats/thieves/cyborgs/orangutans; grammaticality preserved → same category.
    • Semantic oddities (e.g. “ferocious cars”) do not undermine grammatical distribution evidence.

Constituency: Fundamental Notions

  • Constituent: any group of words that grammar treats as a single unit.
  • Phrase: a constituent that possesses a head determining its category and licensing dependents.
    • E.g. “the big dog” = NP → head dog.
  • Constituents nest recursively; phrases contain phrases, yielding hierarchical structure.

Four Core Constituency Tests

  1. Movement (a metaphor)
    • If a sequence can relocate as a block (or surface in parallel constructions) while preserving grammaticality, it is a constituent.
    • English diagnostics:
      • Topicalisation: “At that strange little shop, Nimra bought a scarf.”
      • Passive alternation: “That lovely scarf was bought by Nimra.”
      • Clefts: “It was a book about linguistics that he read.”
    • Incomplete movement (leaving word(s) behind) → ungrammatical.
  2. Substitution / Replacement
    • A constituent can be exchanged for a single pro-form without meaning change.
      • NP: it, them, demonstratives.
      • PP: there/then.
      • VP: do so, verb-phrase ellipsis.
    • “He read a book about linguistics → He read it.”
    • “She opened a restaurant and he did so too.”
  3. Standalone / Answer-to-Question
    • If a sequence can be uttered alone (e.g. short answer), it functions as a constituent.
    • “What did Nimra buy?” → “That lovely scarf.”
  4. Coordination
    • Only like constituents (same structural type/function) may be joined by and/or.
    • “We peeled [the potatoes] and [the carrots].” (NP + NP)
    • “We [peeled the potatoes] and [chopped the herbs].” (VP + VP)
    • Caution: coordination can yield false positives; always corroborate with other tests.

Applying the Tests – Sample Sentences

  • “John eats at really fancy restaurants.”
    • Standalone Q-A: “What does John do on weekends?” → “Eat at really fancy restaurants.”
    • Replacement: “John does so too.”
    • Cleft: “It’s eating at really fancy restaurants that John loves to do.”
    • Incomplete variants (e.g. John eats it really fancy restaurants) are ungrammatical → confirms constituency boundary.

Heads & Dependents

  • Head: obligatory element; determines phrase category, selects dependents, dictates agreement/inflection.
    • Removing head destroys phrase (the only on the spaceship – invalid without “dog”).
  • Dependents:
    • Optional; can often be removed without violating grammar (“the dog” vs. “dog”).
    • Type of dependents licensed varies by head (adjective permissible with nouns; adverb not).

Hierarchical & Recursive Structure

  • Example: “The VC eats at fancy restaurants.”
    • NP → [Det the] + [N VC]
    • VP → [V eats] + [PP at fancy restaurants]
    • PP → [P at] + [NP fancy restaurants]
    • NP → [Adj fancy] + [N restaurants]
  • Each constituent nests within a larger one; recursion allows unlimited sentence length/complexity from finite rules.

Practical Advice for Assignments & Data Work

  • When analyzing an unfamiliar language:
    • Form hypotheses on constituency and category membership.
    • Systematically apply multiple tests; one piece of evidence seldom decisive.
    • Account for semantic oddities or pragmatic factors that may skew intuitions.
    • Document both successful and failed test results to support conclusions.
  • In fieldwork/textual analysis:
    • Elicit or search for parallel constructions (movement, clefts, passives, etc.).
    • Use ellipsis and pronoun replacement to probe NP/VP/PP boundaries.

Key Takeaways

  • Grammar groups words into constituents; phrases have heads and dependents.
  • Four core diagnostics—movement, substitution, standalone, coordination—are essential analytic tools.
  • Categories are identified by distribution, not by meaning alone.
  • Evidence is cumulative; the more tests a proposed constituent passes, the more robust the analysis.
  • Understanding constituency (≈ 60\% of the course’s intellectual load) underpins all subsequent syntactic reasoning.