@cite{sharvit-2003}: Simultaneous Tense as a Genuine Tense #
@cite{ogihara-sharvit-2012} @cite{sharvit-2003}
Sharvit's theory: the simultaneous reading under attitudes arises from a genuine tense with its own semantics -- it denotes "at the local evaluation time" -- rather than being zero tense (Ogihara), deleted past (Kratzer), or a bound variable (Abusch).
Core Mechanisms #
- Simultaneous tense: a tense morpheme whose semantics directly locates the event AT the local evaluation time
- Indirect question SOT: the simultaneous tense handles SOT in indirect questions without deletion, zero tense, or binding
- RC tense: uniform mechanism for relative clause tense
- Embedded present puzzle: simultaneous tense at a future eval time explains present-under-future
Motivating Data #
The key data that Sharvit uses to motivate a separate simultaneous tense:
Indirect questions: "John asked who was sick" has a simultaneous reading (who was sick at the asking time). Standard SOT theories struggle here:
- Deletion: what triggers deletion in a question (no attitude verb with the right properties)?
- Zero tense: why is the question past-marked at all?
- Binding: no attitude-like semantics provides a binder in questions.
Sharvit's solution: the "past" in the indirect question IS a simultaneous tense -- its semantics locates the event at the local evaluation time (the asking time), not before it.
Optional SOT #
In Hebrew-type languages, both the simultaneous tense and genuine past are available in embedded clauses, producing optional SOT:
- "John said Mary was sick" (simultaneous tense → simultaneous reading)
- "John said Mary is sick" (present tense → simultaneous reading too, but with different pragmatic properties)
Limitations #
- The formal implementation here uses
sorryfor some derivation theorems pending closer consultation of Sharvit's LI 2003 paper - Cross-linguistic predictions not fully worked out
Sharvit's simultaneous tense: a genuine tense morpheme whose semantics locates the event AT the local evaluation time.
This is distinct from:
- Zero tense (Ogihara): an ambiguous reading of past
- SOT deletion (Kratzer): removal of embedded past
- Bound variable (Abusch): a variable receiving the matrix event time
The simultaneous tense is a first-class tense with its own denotation: ⟦SIM⟧ = λt. λe. time(e) = t (event time = evaluation time)
- evalTime : Time
The local evaluation time (set by the matrix verb)
- eventTime : Time
The event time
The tense's semantics: event time = evaluation time
Instances For
The simultaneous tense is not a past tense: it does not assert temporal precedence. It asserts temporal overlap.
Sharvit's challenge: tense in indirect questions. "John asked who was sick" -- past in the indirect question gets a simultaneous reading. Standard SOT theories struggle:
- Deletion: what triggers deletion in a question?
- Zero tense: why is the question past-marked at all?
- Binding: no attitude-like semantics in questions. Sharvit: the simultaneous tense has its own semantics.
- askingTime : Time
The asking time
- sickTime : Time
The sickness time
Simultaneous: sickness at the asking time
Instances For
An indirect question with simultaneous tense has the embedded event at the matrix event time (asking time), just like declarative SOT.
The embedded present puzzle: "John will say Mary is sick."
The present tense "is" under future "will say" gets a simultaneous reading: Mary is sick at the future saying time, not necessarily at speech time.
Sharvit's account: the "present" in the embedded clause is the simultaneous tense, evaluated at the (future) saying time.
- speechTime : Time
Speech time
- sayingTime : Time
The future saying time
- sickTime : Time
The sickness time
- future_saying : Prop
The saying is in the future
Simultaneous tense: sick at the saying time
Instances For
Under Sharvit's account, the embedded present puzzle is resolved: the simultaneous tense at a future eval time correctly predicts that sickness is at the future saying time.
Optional SOT in Hebrew-type languages.
Unlike English (obligatory SOT), Hebrew allows both:
- "John said Mary was sick" (simultaneous tense → simultaneous reading)
- "John said Mary is sick" (present tense → also simultaneous, but with double-access pragmatics)
The availability of both forms with similar (but not identical) meanings supports the existence of a dedicated simultaneous tense: in Hebrew, the choice is not forced by grammar, but in English, SOT forces the "past" form even for the simultaneous reading.
- pastFormSimultaneous : HebrewSOTChoice
Past morphology → simultaneous tense (Sharvit)
- presentFormDoubleAccess : HebrewSOTChoice
Present morphology → genuine present (double-access)
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Both Hebrew forms are grammatical (optional SOT).
Sharvit's account of RC tense: the simultaneous tense in a relative clause is evaluated at the modified NP's temporal coordinate.
"the man who was tall" — the past in the RC is simultaneous tense evaluated at the temporal coordinate of "the man."
This provides a uniform mechanism: the same simultaneous tense operates in attitude complements, indirect questions, and RCs.
- npTime : Time
The modified NP's temporal coordinate
- rcEventTime : Time
The RC event time
Simultaneous: RC event at the NP's time
Instances For
RC simultaneous tense gives the expected frame: R = P.
Sharvit derives the simultaneous reading directly: the simultaneous tense semantics gives R = P by definition.
Sharvit derives the indirect question simultaneous reading without deletion, zero tense, or binding. The simultaneous tense in "John asked who was sick" locates sickness at asking time.
Sharvit handles the embedded present puzzle: "John will say Mary is sick" → sick at saying time. The simultaneous tense at a future eval time does the work.
Sharvit derives RC tense uniformly via the simultaneous tense mechanism (same tense, different evaluation contexts).