Quantifier Interaction and Multiple Quantification

Core Idea

When multiple quantifiers appear in a sentence, their relative scope determines meaning. Quantifier interactions create specific patterns of readings constrained by syntax and pragmatics, not all logically possible scope orderings are available.

How It's Best Learned

Begin with two-quantifier sentences like 'Every student read some book' and systematically map scope orderings, then identify which readings are available and why.

Explainer

From your study of first-order logic, you know that quantifiers like ∀ (for all) and ∃ (there exists) bind variables and give sentences their logical form. You also know from your study of quantifier scope ambiguity that the *order* of quantifiers matters: "∀x ∃y loves(x,y)" (everyone loves someone) says something different from "∃y ∀x loves(x,y)" (there is someone whom everyone loves). The second reading posits a single beloved person; the first allows a different beloved for each person. When a natural language sentence contains multiple quantifiers, the question of which interpretation is expressed — and which interpretations are even *available* — is the problem of quantifier scope interaction.

The canonical example is "Every student read some book." In first-order logic, this has two possible formalizations. The surface scope reading gives the universal quantifier wide scope: ∀x [student(x) → ∃y [book(y) ∧ read(x,y)]] — for every student, there is some book they read, and different students may have read different books. The inverse scope reading gives the existential quantifier wide scope: ∃y [book(y) ∧ ∀x [student(x) → read(x,y)]] — there is a specific book that every student read. In English, the surface scope reading is strongly preferred, but the inverse scope reading is also available in context. What constrains which readings are available?

Syntax plays a critical role. The theory of Quantifier Raising (QR) — developed within generative linguistics — proposes that quantifiers are covertly moved to a position of c-command over their scope at a level of syntactic representation called Logical Form (LF). The position a quantifier occupies in the surface syntactic structure tends to determine its "default" scope, which is why surface scope readings are easier. Inverse scope readings require a more complex derivation in which a quantifier has been raised across another. This predicts asymmetries: inverse scope becomes harder or unavailable with certain syntactic constructions, explaining the data cross-linguistically.

Inverse linking is a revealing special case. In "Every student in some university passed," the universal "every student" contains an indefinite "some university." Despite the nested structure, the indefinite can take wide scope over the universal: "There is some university such that every student in it passed." This inverse reading is available even though "some university" is syntactically embedded inside the subject DP. QR handles this by allowing "some university" to raise out of its containing phrase to a higher LF position. The availability of inverse linking is a strong empirical argument that scope is determined at an abstract level of logical form, not simply read off from surface word order.

The deepest issue is scope rigidity vs. scope freedom. Some languages, like Chinese and Hindi, are argued to be more "scope rigid" — they tend to have only surface scope readings, and inverse scope requires special morphology or prosody. Other languages, like English and German, allow more scope freedom. This cross-linguistic variation suggests that the constraints on quantifier scope are not purely logical but involve the specific grammatical resources of a language. For the philosopher of language, this raises a fundamental question: is logical form a universal structure that all natural languages express differently, or is the "logical form" of a natural language sentence a product of its specific grammar? The behavior of quantifier scope interactions provides some of the sharpest evidence bearing on this question.