Zheng (2025): Nandao-Q Felicity @cite{zheng-2025} #
Mandarin nandao-question felicity. Self-contained study file:
empirical data, the Mandarin Fragment entry, and bridges to the
Kernel-theoretic felicity predicate nandaoFelicitous.
The core finding is that positive evidential bias is necessary for nandao-Q felicity, while negative epistemic bias is neither necessary nor sufficient.
Key Generalizations #
- Positive evidential bias (contextual evidence for p) → nandao felicitous
- Epistemic bias alone (prior belief against p, no evidence) → nandao infelicitous
- Evidence must be unexpected relative to prior information state
- Nandao-Qs can function as pure inquiry (no prior belief required)
Predictions verified #
fragment_data_evidential: Fragment entry's evidential bias requirement matches the empirical generalizationfragment_data_epistemic: Fragment entry correctly does not require epistemic biaskernel_requires_evidence: KernelnandaoFelicitousentailsevidenceSupportsnandaoFullFelicity: integrated two-layer felicity (singleton sister presupposition ∧ Kernel-bias check) — §5biasedUse_integrated_felicity: dripping-raincoat scenario satisfies integrated felicity at the §5 level — §6biasedUse_witnesses_integrated_felicity: §1 datum (ex. 2) ↔ §6 theoretical prediction
Known gaps #
- No formalization of the unexpectedness requirement in the Kernel theory
A nandao-Q felicity datum.
- exampleNum : String
Example number from @cite{zheng-2025}
- context : String
Context description
- sentence : String
The nandao-Q sentence (pinyin)
- evidentialBias : Bool
Is there positive evidential bias (contextual evidence for p)?
- epistemicBias : Bool
Is there negative epistemic bias (prior belief against p)?
- unexpectedEvidence : Bool
Is the evidence unexpected?
- felicitous : Bool
Is the nandao-Q felicitous?
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Ex. 1: Rhetorical use. Lee working on Sunday (evidence) contradicts B's norm that people don't work Sundays (epistemic/deontic bias).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Ex. 2: Biased question. A believes not-raining; B enters with dripping raincoat (evidence contradicting belief).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Ex. 3: Pure inquiry (NOVEL). Same evidence as ex. 2 but A has NO prior belief about the weather. Nandao is still felicitous.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Ex. 4a: Epistemic bias without evidence. Speaker believes room is empty but has no contextual evidence either way.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Ex. 5 ctx 1: Evidence + no belief → felicitous.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Ex. 5 ctx 2: No evidence + no belief → infelicitous.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Ex. 5 ctx 3: No evidence + epistemic bias → infelicitous.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Ex. 6 ctx 1: Unexpected evidence → felicitous.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Ex. 6 ctx 2: Expected evidence → infelicitous.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Generalization 1: All felicitous nandao-Qs have evidential bias.
Generalization 2: Some felicitous nandao-Qs lack epistemic bias (the pure inquiry use).
Generalization 3: Some infelicitous nandao-Qs have epistemic bias (epistemic bias is not sufficient).
Generalization 4: All felicitous nandao-Qs have unexpected evidence.
9 data points from 6 examples covering 4 conditions.
The nandao Fragment entry's evidential bias requirement matches the empirical generalization: all felicitous nandao-Qs have evidential bias.
The nandao Fragment entry correctly does NOT require epistemic bias, matching the empirical finding that some felicitous nandao-Qs lack it.
Kernel nandaoFelicitous entails evidenceSupports, connecting the
Theory predicate to the Fragment's requiresEvidentialBias = true and
the empirical generalization evidential_bias_necessary.
Zheng's layer assignments for the three Mandarin Q-particles in the
@cite{dayal-2025} cartography [SAP [PerspP [CP ...]]]. The _
argument is unused: the layer is a theoretical overlay on the
fragment particle, not a computed property of its lexical fields.
Instances For
ma is the unmarked CP-layer particle: widest distribution (matrix, subordinated, quasi-subordinated).
ba and nandao are PerspP-layer biased particles: matrix + quasi-subordinated only.
The layer split mirrors the bias-profile split: the unbiased particle is CP, the biased ones are PerspP.
@cite{bhatt-dayal-2020} fn. 11 explicitly cites the parallel
Mandarin nandao analysis as the model for their kya: proposal.
At the algebraic level, both particles share the same singleton
presupposition: their sister question must denote a singleton-cell
issue (@cite{bhatt-dayal-2020} eq. 23). The cross-particle
generalization is captured by inheriting the same
Core.Question.IsSingleton predicate — both kya: and nandao take a
SingletonQuestion W as well-typed sister content.
nandao is felicitous on a one-cell ("highlighted") polar — same
canonical good-input case as kya:. The proof reduces to
isSingleton_declarative, identical to the kya: side.
Bridge §2 and §4. Nandao's full felicity has two independent layers:
- Layer 1 (Question-level, §4): the sister content
Q : Core.Question Worldis singleton —alt Q = {p}for some unique witnessp. This is the @cite{bhatt-dayal-2020} eq. 23 presupposition: nandao requires a one-cell sister, not a two-cell Hamblin polar. - Layer 2 (Kernel-level, §2): the Kernel-bias check
nandaoFelicitous k u pholds for the witness — the evidence inKraisesP(p),Kis incompatible with the priorU, andpis not directly settled.
These layers are independent concerns: Layer 1 is about semantic
well-formedness of the sister content, Layer 2 is about discourse
felicity in context. The integrated predicate composes them into a
single statement, and the bridges below show that Layer 1 failure
(e.g. a non-trivial two-cell polar) blocks felicity at the integrated
level regardless of (k, u).
Integrated nandao felicity: the conjunction of Layer 1
(singleton presupposition: alt Q = {p}) and Layer 2 (Kernel-bias
check on the witness). The witness p is supplied externally so
decidability is concrete; for the noncomputable choice from a
SingletonQuestion use SingletonQuestion.witness.
Equations
- Phenomena.Questions.Studies.Zheng2025.nandaoFullFelicity Q k u p = (Q.alt = {p} ∧ Semantics.Modality.nandaoFelicitous k u p)
Instances For
Layer-1 projection: integrated felicity entails the §4 singleton presupposition.
Layer-2 projection: integrated felicity entails the §2 Kernel-bias check on the witness.
Layer-1 obstruction: a two-cell Hamblin polar polar p₀ (with
non-trivial p₀) admits no integrated-felicity witness. No
Kernel + Background can rescue it: the §4 type-level barrier
propagates upward through the alt Q = {p} requirement. The
structural reason polar two-cell questions are universally
blocked from nandao licensing.
Declarative reduction: on a one-cell sister declarative p,
integrated felicity is exactly the §2 Kernel-bias check on p.
The Layer-1 component holds trivially because alt (declarative p) = {p} (alt_declarative). This makes the §2 ↔ §5 connection
explicit on the canonical felicitous case.
Apply §5 to the canonical biased-use scenario from @cite{zheng-2025} ex. 2:
- K =
[wearingRaincoat]— direct evidence (B enters with dripping coat) - U =
[expectDry]— A's prior expectation it is not raining - p =
isRaining— the question content
The Kernel-side bias check is raincoat_nandao_felicitous (proven in
Theories/Semantics/Modality/Kernel.lean from explicit cardinality
counts on World4). Pairing it with the singleton presupposition
yields integrated felicity at the §5 level. The §1 datum biasedUse
records the same scenario as empirical data (evidentialBias = true,
epistemicBias = true, felicitous = true); the bridge below makes
the data ↔ theory correspondence explicit.
§1 ex. 2 ↔ §5 integrated felicity: in the dripping-raincoat
scenario with sister declarative isRaining, both layers of nandao
felicity hold simultaneously. Reduces to raincoat_nandao_felicitous
via nandaoFullFelicity_declarative_iff.
Data ↔ theory bridge: the §1 datum biasedUse is felicitous and
has both bias profiles set; the §5 integrated felicity holds for the
matching Kernel scenario. The shared scaffold is the dripping-raincoat
setup of @cite{zheng-2025} ex. 2 — empirical observation on the data
side, derived prediction on the theory side.