Multiplicity Inferences: Empirical Data #
Theory-neutral empirical patterns for multiplicity inferences — the observation that bare plurals trigger a "more than one" reading in upward-entailing contexts but not in downward-entailing contexts.
The Puzzle #
- "Emily fed giraffes" → Emily fed more than one giraffe
- "Emily didn't feed giraffes" ≠ Emily didn't feed more than one giraffe (rather: Emily didn't feed any giraffes)
This monotonicity sensitivity parallels classical scalar implicatures (e.g., "some" → "not all" in UE but not DE contexts).
Theoretical Approaches #
Three main accounts:
- Ambiguity (@cite{farkas-de-swart-2010}): Plural is ambiguous (inclusive "one or more" vs exclusive "more than one"), resolved by Strongest Meaning Hypothesis.
- Implicature (@cite{sauerland-2003}, @cite{spector-2007}, @cite{zweig-2009}): Plural literally means "one or more," multiplicity arises as a scalar implicature with the singular as alternative.
- Homogeneity (@cite{kriz-2015}): Plural interpretation via homogeneity presupposition.
Key References #
- @cite{sauerland-2003}
- @cite{spector-2007}
- @cite{zweig-2009}
- @cite{farkas-de-swart-2010}
- @cite{tieu-etal-2020}
A multiplicity inference datum: a bare plural sentence tested in upward-entailing (positive) and downward-entailing (negative) contexts.
- positiveSentence : String
The bare plural sentence (positive form)
- negativeSentence : String
The negated form
- multiplicityInPositive : Bool
Does the "more than one" inference arise in the positive?
- multiplicityInNegative : Bool
Does the "more than one" inference arise in the negative?
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Core example: "Emily fed giraffes."
In UE: interpreted as "Emily fed more than one giraffe." In DE: "Emily didn't feed giraffes" ≈ "Emily didn't feed any giraffes."
Equations
- One or more equations did not get rendered due to their size.
Instances For
Conditional antecedent (DE context).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Multiplicity arises in UE but not DE for "fed giraffes".
Multiplicity arises in UE but not DE for "books on desk".
The monotonicity sensitivity of multiplicity inferences parallels that of classical scalar implicatures. This structure captures the parallel.
- weakTerm : String
The scalar term (e.g., "some", bare plural)
- strongAlternative : String
Its stronger alternative (e.g., "all", singular)
- inferenceInUE : String
Inference in UE context
- arisesInUE : Bool
Does inference arise in UE?
- arisesInDE : Bool
Does inference arise in DE?
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Some/all parallel.
Equations
- Phenomena.Plurals.Multiplicity.someAllParallel = { weakTerm := "some", strongAlternative := "all", inferenceInUE := "not all", arisesInUE := true, arisesInDE := false }
Instances For
Plural/singular parallel.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Or/and parallel.
Equations
- Phenomena.Plurals.Multiplicity.orAndParallel = { weakTerm := "or", strongAlternative := "and", inferenceInUE := "not both", arisesInUE := true, arisesInDE := false }
Instances For
Some/all: implicature in UE, not in DE.
Plural/singular: multiplicity in UE, not in DE.
Or/and: exclusivity in UE, not in DE.
Competing theoretical approaches to multiplicity inferences.
- ambiguity : PluralTheory
Plural is ambiguous; Strongest Meaning Hypothesis resolves.
- implicature : PluralTheory
Plural literally means "one or more"; multiplicity is implicature.
- homogeneity : PluralTheory
Plural interpretation via homogeneity presupposition.
Instances For
Equations
- Phenomena.Plurals.Multiplicity.instDecidableEqPluralTheory x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
The fundamental discriminating property: does the theory analyze multiplicity as arising via the same mechanism as scalar implicatures?
This is the single primitive from which all empirical predictions are derived. The implicature theory says multiplicity IS an SI; ambiguity says it arises from lexical polysemy + Strongest Meaning; homogeneity says it arises from presupposition.
Equations
Instances For
Children undercompute SIs (@cite{noveck-2001}). If multiplicity IS an SI, children should compute fewer multiplicity inferences.
Equations
Instances For
SIs sharing a mechanism have correlated rates within individuals (Uniformity Prediction). Multiplicity rates should correlate with standard SI rates iff they share the SI mechanism.
Equations
Instances For
SIs show UE/DE polarity asymmetry. If multiplicity is an SI, the polarity asymmetry in children follows from children's general difficulty with SIs.
Equations
Instances For
In singular contexts (exactly one object acted on):
- If multiplicity is an SI: positive is literally true + false implicature (misleading); negative is literally false → different truth-value status.
- If multiplicity is lexical/presuppositional: both are undefined or both false → same status. Only the SI mechanism predicts asymmetric truth-value judgments.
Equations
Instances For
All four predictions reduce to the single mechanistic property. This makes the structure explicit: we don't have four independent stipulations, but one property with four consequences.
The implicature theory is uniquely identified by ANY of the four
predictions (since they all reduce to usesSIMechanism).
Any single prediction suffices to identify the implicature theory,
since all predictions reduce to usesSIMechanism. Here we show this
for predictsChildrenComputeFewer (the others are identical).
Singular context asymmetry identifies implicature.
The competing theories do NOT use the SI mechanism.