Docs / Studies / Studies.BarLevBassi2026

Infelicity with homogeneity: focus, not (necessarily) presupposition

Bar-Lev, M. E. & Bassi, I. (2026). Journal of Semantics, 43(2), 139–165

Guerrini and Wehbe derive the oddness of she read the books after I knew she read some of them from a homogeneity presupposition that leaves the assertion uninformative. That verdict depends on the context alone, so it extends to the acceptable she DID read the books. Bar-Lev and Bassi blame focus: the least F-marking of the clause congruent to a background question marks one head, the covert domain variable for some or all? and the polarity head for all or not all?, and a plural definite is acceptable where that head is pronounced. A homogeneity presupposition, where one is needed, is that of the polar question the sentence answers, and accommodating it gives their sentence.

Implementation notes

  • The clause is the implicature account's, Exh over Bar-Lev's existential pluralization, with two heads that can be F-marked: the polarity head, above Exh, and the domain variable of the pluralization, whose alternatives are its subdomains. Alternatives to the overt lexical material are not modelled; the paper assumes with Fox and Katzir that they are no more complex than the material and so are not quantificational.
  • Congruence to some or all? is the appendix's, that each answer be a disjunction of focus alternatives, Q ⊆ supClosure fv; under the paper's containment Q ⊆ fv no marking is congruent. Focus minimization is Minimal over the markings, ordered by which heads are F-marked.
  • The atoms are an independent family: every set of them is the set of those satisfying the predicate at some world.
  • A row holds a plural definite and, as its alternatives, the quantified counterparts the paper judges in the same context. The word printed in capitals in an acceptable counterpart shows the background question, by which fragment entry it is: a quantifier or an auxiliary. This is the paper's method of taking the stress of the overt quantifier as the guide to the question.

TODO

  • The alternative on which the focus value keeps the alternatives without Exh as well, and the accommodation of a question other than the one asked (§5.3).
  • The question none or all? of (13a-ii) has the answers of the polar question (43a) and still rejects the plural answer; the paper takes congruence to see the form of the question, which a set of propositions does not record.
  • Focus on overt material inside the plural, his PHONOLOGY papers, and the legality question behind she's not MARRIED to those people.
  • The clause is affirmative; the negative rows are read by the same map from stress to head.
  • Local accommodation licensed by stress on a focus associate of pex (§7), and the other homogeneity phenomena of §8.

References

The homogeneity presupposition and informativity (§2, §3)

With the homogeneity presupposition a plural predication is uninformative in every context that entails that some atom satisfies the predicate. Focus plays no part, so she DID read the books, (16), and he worked on his PHONOLOGY papers, (39a), are predicted to be as odd as she read the books, (3a), in the same contexts.

theorem not_informative_of_subsetL84–92
{Atom : Type u_1} {W : Type u_2} [DecidableEq Atom] {x : Finset Atom} {P : Atom → W → Prop} {R : Set (Set W)} {C : Set W} (hC : C ⊆ ⋃ a ∈ x, BarLev2021.atom P a) :

The focus alternatives of a plural predication (§5, appendix A)

L116
L134–138
L141–144
L152–158

Whatever is F-marked, the existential answer is not among the focus alternatives of the clause, so the clause is not congruent to the question whether Ann read some or all of the books, (33), when congruence is containment.

theorem biUnion_notMem_alternatives_clauseL195–210
{Atom : Type u_1} {W : Type u_2} {x : Finset Atom} {P : Atom → W → Prop} (h : Exhaustification.IsIndependentFamily x (BarLev2021.atom P)) (hcard : 2 ≤ x.card) (m : Head → Focus.Mark) :
⋃ a ∈ x, BarLev2021.atom P a ∉ (clause x P m).alternatives

Built from ,

The least F-marking of the clause congruent to a background question marks one head and nothing else: the domain variable for some or all?, the polarity head for all or not all?, with the homogeneity inference computed below it.

theorem minimal_congruent_iffL268–273
{Atom : Type u_1} {W : Type u_2} {x : Finset Atom} {P : Atom → W → Prop} (h : Exhaustification.IsIndependentFamily x (BarLev2021.atom P)) (hcard : 2 ≤ x.card) (q : Background) {m : Head → Focus.Mark} :
Minimal (Background.Congruent x P q) m ↔ ∀ (i : Head), m i = Focus.Mark.focused ↔ i = q.head

Built from , , ,

The homogeneity-focus generalization (§4)

A plural definite is odd where its quantified counterpart needs stress on the quantifier. The stress on the counterpart shows the background question, and the plural definite is acceptable when the head that question asks to be F-marked is pronounced.

L289–292
L295–302

The homogeneity-focus generalization (18). A row holds a plural definite and, as its alternatives, the quantified counterparts judged in the same context. Where an acceptable counterpart shows a background question, the plural definite is acceptable exactly when the head that question asks to be F-marked is pronounced: it is odd in (3), (9), (11), (13), (15), (19), (20) and (38), where ALL or ANY bears the stress, and fine in (16) and (37), where DID does.

theorem judgmentsL313–322
(e : Datum) :
e ∈ Examples.all → HasDefinite e → ∀ a ∈ e.contrasts, a.2 = Judgment.acceptable → ∀ s ∈ stressed? a.1, ∀ q ∈ background? s, e.judgment = Judgment.acceptable ↔ q.head.Overt

Built from , , , ,

The presupposition of the question (§6)

With the homogeneity inference computed above the polarity head, the polarity alternatives of the clause are the universal reading (BarLev2021.exhIEII_eq_universal) and, exhaustification being vacuous under negation (Exhaustification.exhIEII_compl_biUnion), the negated existential: the question (43a) behind John is not married to his teammates. It presupposes homogeneity, where the question whether all or not all presupposes nothing (Question.presupposition_pair_compl).

A question presupposes that one of its answers is true. For the question whether all or none of the atoms satisfy the predicate that is homogeneity, and the universal answer with this presupposition accommodated is Guerrini and Wehbe's plural predication: the oddness of (41a) and (42a) follows as on their account, from the question and not from the sentence.

theorem pexPlural_eq_condEventL351–375
{Atom : Type u_1} {W : Type u_2} {x : Finset Atom} {P : Atom → W → Prop} [DecidableEq Atom] (h : Exhaustification.IsIndependentFamily x (BarLev2021.atom P)) (hcard : 2 ≤ x.card) :
GuerriniWehbe2025.pexPlural x x P (supClosure (BarLev2021.atom P '' ↑x)) = Trivalent.condEvent (fun (x_1 : W) => x_1 ∈ Question.presupposition {⋂ a ∈ x, BarLev2021.atom P a, (⋃ a ∈ x, BarLev2021.atom P a)ᶜ}) fun (x_1 : W) => x_1 ∈ ⋂ a ∈ x, BarLev2021.atom P a