[AK98]: Tense in Temporal Adjunct Clauses #
Arregui & Kusumoto (SALT VIII, 1998) argue that tense in temporal adjunct clauses (TACs) is not semantically in the scope of the matrix tense. They reject the relative-tense analysis of Ogihara (1994, 1996) on the basis of:
- Japanese when-clauses with past tense (ex 8) are episodic and
acceptable, contradicting a relative-tense prediction of clash with
toki('when'). - Polish before- or after-clauses pattern with English (past + past), but Polish lacks the SOT rule — so the English pattern can't be a SOT-deletion effect.
- Geis-style ambiguities (ex 14) in English when-, before-, after-clauses indicate a relative-clause structure where the embedded tense is evaluated independently with the speech time as its reference.
A&K propose: English and Polish TACs are relative-clause-like (absolute tenses); Japanese before- or after-clauses involve direct TP-selection by the temporal connective (no relative-clause structure); Japanese when-clauses are absolute (like English when). The past-vs-present contrast in Japanese TACs is quantificational (episodic vs habitual), not temporal — present tense is a temporal variable bound by adverbs of quantification.
Empirical anchors (verified vs PDF) #
- (1) "Bernhard said that Junko was sick" — English SOT context-setter.
- (5a/b) Japanese before+past * vs before+present OK.
- (6a/b) Japanese after+past OK vs after+present *.
- (7a/b) English Elliott/Eva with both past+past tense.
- (8) "Junko was in her room when Satoshi came" — Japanese when+past (episodic), the counterexample to Ogihara's relative tense.
- (9) "...whenever Satoshi came" — Japanese when+present (habitual).
- (14) "I encountered Satoshi in Amherst when you said he had left" — Geis ambiguity.
- (18a/b) "I watered the plant before/after it died" — before/after veridicality contrast.
Scope of the Reichenbach frames below #
A&K's English-side anchor example is Elliott left before Eva came
(ex 7a); frames are named elliottLeft / evaCame accordingly.
The (R,E)-frames only capture temporal ordering between two past events. A&K's actual contribution — the relative-clause analysis of English/Polish TACs vs direct TP-selection in Japanese, plus the past/present quantificational contrast — is not encodable in (S,P,R,E) and lives in the JSON above + the verified Geis-ambiguity example.
These frames represent A&K example (7a) — Elliott left before Eva came. The (R,E)-frame only encodes the temporal ordering between
the two past events; A&K's actual claims about TAC structure
(relative-clause vs direct TP-selection) and the
episodic/habitual quantificational contrast are not encoded in
Reichenbach frames. The JSON above carries those facts.
Adjunct clause "before Eva came" — Eva's coming event. Past tense; in A&K's analysis, this past tense in English is interpreted absolutely (with speech time as reference), unlike the Ogihara relative-tense story.
Equations
- ArreguiKusumoto1998.evaCame = { speechTime := 0, perspectiveTime := 0, referenceTime := -3, eventTime := -3 }
Instances For
Matrix clause "Elliott left" — Elliott's leaving event. Past tense; absolute perspective.
Equations
- ArreguiKusumoto1998.elliottLeft = { speechTime := 0, perspectiveTime := 0, referenceTime := -2, eventTime := -2 }
Instances For
before-ordering: the adjunct event precedes the matrix event.
This is the structural consequence of before in the temporal
connective, not of any tense composition.