past possibility crosslinguistically - uni-tuebingen.de possibility crosslinguistically ... –...

17
Universität Potsdam, Sonderforschungsbereich 632 May 21, 2013 Past Possibility Crosslinguistically Vera Hohaus Eberhard Karls Universität Tübingen Plot 1 Introduction and Background 1 2 Starting Point 3 3 Architecture at Logical Form 7 4 Crosslinguistic Evidence 8 5 Concluding Remarks 13 I dwell in Possibility– A fairer House than Prose– More numerous of Windows– Superior – for Doors– Of Chambers as the Cedars– Impregnable of Eye– And for an Everlasting Roof The Gambrels of the Sky– Of Visitors – the fairest– For Occupation – This– The spreading wide my narrow Hands To gather Paradise– (Emily Dickinson) 1. What is the LF architecture of tense and modality? 2. Do modals contribute to or restrict the temporal interpretation of a sentence? 1 Introduction and Background Key properties of standard modal analyses. – Modals are quantifiers over possible worlds. Possibility modals contribute existential, necessity modals universal quantification. The domain of quantification is restricted via an accessibil- ity relation and an ordering source, both contextually provided and subject to certain lexical restrictions (cf. e.g. Kratzer 1981, 1991). Figure 1: Domain Restriction via Accessibility Relation and Ordering Source This is joint work with Hotze Rullmann, Lisa Matthewson, Sihwei Chen, Becky Laturnus, Meagan Louie, Ori Simchen, Claire Turner, and Jozina Vander Klok (University of British Columbia, Vancouver). 1

Upload: vuongduong

Post on 27-Apr-2018

233 views

Category:

Documents


6 download

TRANSCRIPT

Page 1: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

Universität Potsdam, Sonderforschungsbereich 632

May 21, 2013

Past Possibility Crosslinguistically

Vera HohausEberhard Karls Universität Tübingen⇤

Plot

1 Introduction and Background 1

2 Starting Point 3

3 Architecture at Logical Form 7

4 Crosslinguistic Evidence 8

5 Concluding Remarks 13

I dwell in Possibility–

A fairer House than Prose–More numerous of Windows–

Superior – for Doors–

Of Chambers as the Cedars–Impregnable of Eye–

And for an Everlasting RoofThe Gambrels of the Sky–

Of Visitors – the fairest–For Occupation – This–

The spreading wide my narrow HandsTo gather Paradise–

(Emily Dickinson)

1. What is the LF architecture of tense and modality?2. Do modals contribute to or restrict the temporal interpretation of a sentence?

1 Introduction and Background

Key properties of standard modal analyses.

– Modals are quantifiers over possible worlds. Possibility modals contribute existential, necessitymodals universal quantification. The domain of quantification is restricted via an accessibil-ity relation and an ordering source, both contextually provided and subject to certain lexicalrestrictions (cf. e.g. Kratzer 1981, 1991).

Figure 1: Domain Restriction via Accessibility Relation and Ordering Source

⇤ This is joint work with Hotze Rullmann, Lisa Matthewson, Sihwei Chen, Becky Laturnus, Meagan Louie, Ori Simchen,Claire Turner, and Jozina Vander Klok (University of British Columbia, Vancouver).

1

Page 2: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

– Focus will mainly be on possibility modals.For ease of presentation, we will neglect the contribution of the ordering source.1

(1) Jmust K = �whsi.�phs,ti. 8w0 [Rw(w0) ! p(w0)]

(2) J can K = �whsi.�phs,ti. 9w0 [Rw(w0) & p(w0)]

(3) Borussia can win.

(4)�w

canhs,hhs,ti,tii w

hs, ti

�w

0

Borussia he, ti

winhs,he,tii w

0

(5) Jwin K = �whsi.�xhei. winw(x)

(6) 9w0 [Rw@(w0) & winw0(Borussia)]

– Possibility modals can either combine with an epistemic accessibility relation or, in the case ofroot modals, with a circumstantial accessibility relation.

(7) Context:Suppose I acquire a piece of land in a far away country and discover that soil and climateare very much like at home, where hydrangeas prosper everywhere. Since hydrangeas are myfavorite plants, I wonder whether they would grow in this place and inquire about it.

(8) Hydrangeas can grow here.(Kratzer 1991, p. 646, ex. (21a))

(9) Context:Suppose now that the country we are in has never had any contacts whatsoever with Asia orAmerica, and the vegetation is altogether different from ours.

(10) #There might be hydrangeas growing here.(Kratzer 1991, p. 646, ex. (21b))

However, possibilities and circumstances change over time. A proposition, a set of possibleworlds, might be contained in the set of worlds confirming with what one knows or withwhat the facts are at one point in time but not at another.

Questions to be answered from a crosslinguistic perspective:1. What is the LF architecture of tense and modality?

2. Do modals contribute to or restrict the temporal interpretation of a sentence?3. Does the grammar allow us to evaluate epistemic possibilities in the past?

Preview of answers from currently 15 languages from ten language families:2

2. No.3. Yes.

1 Semantic types are hei, hti, hsi, hvi, hii. If � and ⌧ are types, then h�, ⌧i is a type. Nothing else is a type. Semanticdenotation domains are hei for individuals, hti for truth values, hsi for worlds, hvi for events, and hii for times.

2 Data collection is still in progress.

2

Page 3: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

Language Family Endangered?English GermanicDutch GermanicHebrew SemiticMandarin Sino-TibetanSt’at’imcets Salish xSenćoten Salish xHul’q’umi’num’ Salish xGitxsanimx Tsimshianic xBlackfoot Algonquian xAtayal Formosan (Austronesian) xPaiwan Formosan (Austronesian) xPaciran Javanese Western Malayo-Polynesian (Austronesian)Samoan Polynesian (Austronesian)Ktunaxa Isolate x

Table 1: Current Language Sample and Language Families

Figure 2: Geographical Distribution of Languages

Fieldwork methodology:– acceptability judgment tasks– elicited production tasks and storyboard tasks

(cf. e.g. Matthewson 2004, 2011)

2 Starting Point

Present possibilities.

(11) John might win.

(12) ‘In some world compatible with my current knowledge,John wins the game at some future point.’

3

Page 4: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

Figure 3: Present Temporal, Epistemic Perspective and Future Temporal Orientation

(13) temporal perspective:time interval during which the alternative possible worldsare accessed from the actual world (evaluation time on modal)

(14) temporal orientation:time at which the proposition is true in the accessible worlds(evaluation time on complement of embedded proposition)

strategy 1: lexical strategy 2: decompositionaltemporal orientation and/or perspective temporal orientation and/or perspectivecontributed by the lexical entry contributed by embedded temporal andof the modal itself (e.g. Enç 1996) aspectual operators (e.g. Condoravdi 2002)

Table 2: Multiple Routes to One Meaning

Along with Condoravdi (2002), we are going to pursue a fully decompositional strategy.

Past possibilities (Condoravdi 2002).

(15) John might have won.

(16) John might have won the game (but I’m not sure if he did).‘In some world compatible with my current knowledge,John won the game at some point in the past.’

Figure 4: Present Temporal, Epistemic Perspective and Past Temporal Orientation

(17) John might have won the game (if he hadn’t been sick that day).‘In some world compatible with the facts at a past time,John won the game at some later time in the past.’

Figure 5: Past Temporal, Circumstantial Perspective and Future Temporal Orientation

4

Page 5: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

Figure 6: Available Temporal Configurations according to Condoravdi (2002)

– “Note that there are no modals with a past perspective and a past orientation.“(Condoravdi 2002, p. 63)

– Also, there are no circumstantial readings with present temporal perspectiveand past temporal orientation.

– Most importantly, there are no epistemic modals with past temporal perspective andfuture or past temporal orientation. This is the claim we’re particularly interested in.3

Condoravdi (2002)’s analysis, slightly adopted.

– Deriving the observed pattern of readings involves two components, the particular setup of the ar-chitecture at Logical Form with a lack of covert temporal operators and general, non-grammatical,restrictions regarding the nature of possibility.

(18) Jmight K = �whsi.�thii.�phs,hi,tii. 9w0 [Rw(w0) & p(w0)(t)]

(19) J have K = �whsi.�thii.�phs,hi,tii. 9t0 [t0 < t & p(w)(t0)]Present tense morphology is semantically vacuous.

(20) Jpfv K = �whsi.�thii.�phs,hv,tii. 9e [p(w)(e) & ⌧(e) ✓ [t;_)]

(21) might > have hs, hi, tii

�w

000

�t

00

hhs, hi, tii, ti

mighths, ht, hhs, hi, tii, tiii

w

000t

00

hs, hi, tii

�w

00

�t

0

hhs, hi, tii, ti

havehs, hi, hhs, hi, tii, tiii

w

00t

0

hs, hi, tii

�w

0

�t

hhs, hv, tii, ti

pfv

hs, hi, hhs, hi, tii, tiiiw

0t

hs, hv, tii

�w

�e

Johnrun w

e

3 In that, we’re not alone: Groenendijk and Stokhof (1975), Cinque (1999), Stowell (2004), and Hacquard (2006) e.g.claim that epistemic modals in English do not allow for a past temporal perspective. Eide (2003), Boogart (2007),and Fintel and Gillies (2008), for instance, argue for the existence of such readings.

5

Page 6: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

(22) 9w0 [Rw@, tnow

(w0) & 9t0 [t0 < t

now

& 9e [winw0(e)(John) & ⌧(e) ✓ [t0;_) ]]]‘In some world compatible with my current knowledge,John won the game at some point in the past.’(present temporal perspective, past temporal orientation)

(23) have > might hs, hi, tii

�w

000

�t

00

hhs, hi, tii, ti

havehw, ht, hhs, hi, tii, tiii

w

000

t

00

hs, hi, tii

�w

00

�t

0

hhs, hi, tii, ti

mighths, hi, hhs, hi, tii, tiii

w

00t

0

hs, hi, tii

�w

0

�t

pfv

hs, hi, hhs, hv, tii, tiiiw

0t

hs, hv, tii

�w

�e

Johnrun w

e

(24) 9t0 [t0 < t

now

& 9w0 [Rw@, t0(w0) & winw0,t0(John)]]‘In some world compatible with my knowledge at some past time,John won the game at that or a later point in the past.’(past temporal perspective, future temporal orientation4)

– Aside: This is often construed as a scope ambiguity. Note that, although the respective scoperelationships between might and have differ for the two readings, this structural configuration isnot derived via Quantifier Raising as movement of a quantifier at Logical Form is vacuous andwill thus not result in the desired truth conditions.

– Limiting the inventory of temporal operators to just the visible have, predicts an unavailabilityof readings with a past temporal perspective and a past temporal orientation, regardless ofthe type of accessibility relation, the two lower right boxes in Figure 6.

– Nothing in the setup predicts the unavailability of epistemic possibility modals with a pasttemporal perspective and a future temporal orientation. Also, nothing excludes circumstantialpossibility modals with a present perspective and a past temporal orientation, the pink boxes inFigure 7 below. (These are derived by external factors.)

4 Note that Condoravdi (2002) derives the future orientation from an aspectual, perfective meaning component, repre-sented here by pfv with the meaning in (20). Aspect is, under her analysis, actually incorporated as part of eitherthe lexical entry of have or the modal, which ever sits right above the verb phrase.

6

Page 7: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

Figure 7: Temporal Configurations according to Condoravdi (2002)

– While we, too, favor a decompositional approach, we will show that the empirical predictions ofthe particular setup introduced above are inadequate. In particular, we want to argue for theavailability of epistemic possibility with a past temporal perspective, as in Table 8.5

Figure 8: Readings Suggested to be Actually Available

3 Architecture at Logical Form

– I propose an LF architecture for events, worlds, and times, which is adopted, with only minorchanges, from Stechow and Beck (to appear).6 Notice that rather than having to schlep along thedifferent variables, they are introduced and abstracted over where needed.

(25) Jwin K = �whsi.�ehvi.�xhei. wine,w(x)

(26) Jpfv K = �thii.�phv,ti. 9e [p(e) & ⌧(e) ✓ t]

(27) J impf K = �thii.�phv,ti. 9e [p(e) & ⌧(e) � t]

(28) J have K = �thii.�phi,ti. 9t0 [t0 < t & p(t0)]

(29) Jprosp K = �thii.�phi,ti. 9t0 [t0 � t & p(t0)]

– Perfect have and the silent prospective operator are in complementary distribution with a phrasalprojection of their own. Their semantics is probably slightly more sophisticated, involving anextended now temporal interval (cf. e.g. Alexiadou, Rathert, and Stechow 2003; Stechow 1999).

(30) Jmay K = �whsi.�thii.�phs,ti. 9w0 [Rw,t(w0) & p(w0)]

(31) Jpast K = �thii.�phi,ti. 9t0 [t0 < t & p(t0)]Present tense is semantically vacuous.

5 I’ll also briefly come back to the restrictions on circumstantial modalityin the blue boxes regarding temporal orientation.

6 Focus throughout this talk will be on eventive predicatesbut see the discussion at the end for an extension to stative predicates.

7

Page 8: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

(32) hs, hi, tii

�w

0

�t

00 hti

�pres

past

hi, hhi, ti, tiit

00hi, ti

�t

0

mighths, hi, hhs, ti, tiii

w

0t

0hs, ti

�w

�perf

prosp

hi, hhi, ti, tiit

0hi, ti

�t

�pfv

ipf

hi, hhv, ti, tiit

hv, ti

�e

Johnwin w

e

– This setup allows for sixteen combinations but we will focus only on those eight with perfectiveaspect.7 Crucially, there is nothing in the architecture proper that prevents any type of modalityoccurring with any type of temporal or aspectual operators. Also, we predict the availability ofepistemic modals with past temporal perspective when past binds their temporal argument.

Figure 9: Predicted Temporal Readings for Possibility Modals and their Availability

4 Crosslinguistic Evidence

– Under a null variation null hypothesis (cf. Matthewson 2004), crosslinguistic data can serve asevidence that this are indeed the architecture and the predictions we want.

4.1 Evidence for Temporal Perspective from Tense Operators

– Evidence for temporal operators contributing the temporal interpretation of the modal, its tem-poral perspective, comes from languages in our sample with more transparent tense morphology,such as German and Dutch.

7 If we take the possibility of a future operator above the modal into consideration,we would allow for even more combinations.

8

Page 9: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

—German (Vera Hohaus)—

(33) epistemic possibility, present temporal perspective, past temporal orientation:Der 1. FC Köln könnte das Spiel am vergangenen Sonntag gewonnen haben.(Ich bin mir allerdings nicht ganz sicher.)

Figure 10: Present Temporal, Epistemic Perspective and Past Temporal Orientation

(34) circumstantial possibility, past temporal perspective, future orientation:Lewandowski hätte in der zweiten Halbzeit ein Tor schießen können,(wäre er nicht ausgewechselt worden.)

(35) circumstantial possibility, past temporal perspective, future orientation:Borussia Dortmund konnte zu diesem Zeitpunkt sogar noch gewinnen.

Figure 11: Past Temporal, Circumstantial Perspective and Future Temporal Orientation

(36) hs, hi, tii

�w

0

�t

00 hti

past

hi, hhi, ti, tiit

00hi, ti

�t

0

könnenhs, hi, hhs, ti, tiii

w

0t

0hs, ti

�w

prosp

hi, hhi, ti, tiit

0hi, ti

�t

pfv

hi, hhv, ti, tiit

hv, ti

�e

Dortmund

gewinnen w

e

(37) �ehvi. gewinnenw,e(Dortmund)�thii. 9e [gewinnenw,e(Dortmund) & ⌧(e) ✓ t]�whsi. 9t00 [t00 � t

0 & 9e [gewinnenw,e & ⌧(e) ✓ t

00]]�t

0. 9w00 [Rw0,t0(w00) & 9t00 [t00 � t

0 & 9e [gewinnenw00,e & ⌧(e) ✓ t

00]]]

9

Page 10: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

(38) 9t0 [t0 < t

now

& 9w [Rw@,t0(w) & 9t [t � t

0 & 9e [gewinnenw,e & ⌧(e) ✓ t]]]]

(39) Context:Warum hast du die Einfahrt heute morgen gestreut?

(40) epistemic possibility, past temporal perspective, future orientation:Es hätte schneien können.

(41) Context:My brother blew up a pipe bomb in our mailbox one morning. The neighbor called the cops.Later, my extremely embarrassed mother asked the neighbor why he had called the police. Hereplied:

(42) epistemic possibility, past temporal perspective, past orientation:Es hätte jemand erschossen worden sein können.Es hatte ja zu diesem Zeitpunkt auch jemand erschossen worden sein können.

4.2 Evidence for Temporal Orientation from Perfective and Prospective

– Evidence for second head above AspP contributing to the temporal orientation comes from Ger-man have and Dutch hebben and for prosp from Blackfoot, Gitxsanimx, Ktunaxa, and Mandarin.

—Mandarin (Shiwei Chen)—

Figure 12: 7-Eleven Convenient Stores, Downtown Vancouver

(43) Context:You thought you were going to meet your friend at 41 St.’s 7-11, but you didn’t see him atthe appointed time. You didn’t have a cell with you so you only waited there but never foundhim. Later, when you came home, you got his call, saying: “Why didn’t you go find a boothand call me! I was waiting for you at 44 St.’s 7-11 for an hour!” You reply:

(44) Niyou

kěnéngpossibility(epist.)

huì

prosp

zhao-bú-dàofind-neg.-out

wo.me

‘You might not have found me.’(past temporal, epistemic perspective, future temporal orientation)

10

Page 11: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

(45) Context:You and your friend agreed to meet at 41 St.’s 7-11 but you didn’t see him at the appointedtime. The 7-11 clerk told you there’s another 7-11 on 41 St., so you hastened to go there butstill didn’t find him. When you came home, you got his call. He says: “Why didn’t you waitfor me? I was only 15 minutes late!” You reply:

(46) Niyou

kěnéngpossibility(epist.)

qù-lego-pfv

lìngwàianother

yi-jiaone-cl.

7-11.7-11

‘You might have gone to another 7-11.’(past temporal, epistemic perspective and past temporal orientation)

—Blackfoot (Meagan Louie)—

(47) Context:The FBI has recruited a former art-thief to help them catch other art-thieves. At one point, Ihave no choice but to leave this former conman in a museum (I had to go to the bathroom),when the security precautions are down, in the vicinity of a very famous painting. The conmanseems to be reformed, as he didn’t steal the painting, but my supervisor, who still doesn’t trustthe former conman, berates me, saying:

Figure 13: Inspiration from the TV Series White Collar

(48) Áak-ohkott-ikamo’tsat-yiiprosp-possibility(circ.)-steal

anniskthis

sinaaksin.painting

‘He could have stolen this painting.’(past temporal, circumstantial perspective, future temporal orientation)

4.3 Evidence for Availability of Epistemic and Circumstantial Modal Bases

– Evidence for the availability of both modal bases with a past temporal perspective comes fromBlackfoot, Gitxsanimx, and Ktunaxa.

– In Ktunaxa, for instance, }in can only denote epistemic possibility and ta} only circumstantialpossibility.

– In Gitxsanimx, ima(’a) is unambiguously epistemic but da’ak(hl)xw is unambiguously circum-stantial (cf. also Matthewson to appear; Peterson 2010).

—Ktunaxa (Rebecca Laturnus)—

(49) Context: You stop by the side of the highway en route to the Caribou and notice that thereis no sagebrush growing in this stretch. You ask your friend, can sagebrush grow here? Youare interested in whether the conditions for growing are suitable for sage, not whether any isactually growing.

11

Page 12: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

(50) Ta}possibility(circum.)

itukìiP-nigrow-indic.

Pa.knuqìuúqunaPqa.sagebrush

‘Sagebrush can grow here.’(circumstantial)

(51) Context: Now you want to know if any is actually growing along that stretch of the highway.Your friend says it’s possible but he isn’t sure.

(52) }in

possibility(epist.)itukìiP-nigrow-indic.

Pa.knuqìuúqunaPqa.sagebrush

‘Sagebrush may be growing here.’(epistemic)

(53) Context: You were watching the Canucks and at one point in the first period they were up 2:1.At that point, they might have still won (but they didn’t).

(54) Ta}possibility(circ.)

hukakaP-niwin-indic.

miksanbut

qaneg.

hukakaP-ni.win-indic.

‘They might have won but they didn’t win.’(past temporal, circumstantial, perspective, future orientation)

Figure 14: Past Temporal, Circumstantial Perspective and Future Temporal Orientation

(55) Context: Why did you salt the driveway?

(56) }in

possibility(epist.)mapfv

//cxa}prosp

wa}ink’a}P-ni.snow-indic.

‘It might have started snowing.’(past temporal, epistemic perspective, future orientation)

—Gitxsanimx (Lisa Matthewson)—

(57) Context:You’re up in the Suskwa and notice a burnt patch of forest.You know that huckleberries typically take seed in burnt alpine areas.

(58) Da’akhlxw=hlpossibility(circ.)

dimprosp

limxs=hlgrow(pl.)

maa’yberries

go’osun.here

‘Berries might grow here.’(circumstantial)

(59) #Limx=ima=hlgrow(pl.)=possibility(epist.)

maa’yberries

go’osun.here

‘Berries might be growing here.’(epistemic)

(60) Context:You were watching the Canucks and at one point in the first period they were up 2:1. At thatpoint, they might have still won (but they didn’t in the end).

12

Page 13: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

(61) K’aystill

da’akxw-diitpossibility(circ.)(3pl.)

dimprosp

xsdaa-diit,win(3pl.))

iiand

apemph.

nee=diineg.

xsdaa-diit.win(3pl.)

‘They still could have won but they didn’t win.’(past temporal, circumstantial perspective, future temporal orientation)

(62) Context:When you looked out your window earlier today, the ground was wet, so it looked like it mighthave rained. But you found out later that the sprinkler had been watering the ground.

(63) Yugw=imaa=hlimpf=possibility(epist.)

wisrain

da’awhl.then

‘It have been raining then.’(past temporal, epistemic perspective, past temporal orientation)

5 Concluding Remarks

5.1 The (Intransparent) Case of English

Figure 15: Available Temporal Configurations with Epistemic and Circumstantial Modality

– Circumstantial possibility with past temporal orientation of any kind is unavailability becausepresent facts cannot influence past events. However, this is not a grammatical restriction.

– If we allow for reverse causality and thus for circumstances changing the available possibilitiesat some point prior, circumstantial readings with past perspective and past temporal orientationsignificantly improve, I believe.

(64) Context:Harry Potter, using the time turner, defended the dementor at the lake.

(65) ?So, Harry might have rescued Sirius Black.(present temporal, circumstantial perspective, past temporal orientation)

Figure 16: Past Temporal, Circumstantial Perspective and Past Temporal Orientation

13

Page 14: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

– The availability of both, a vacuous present tense and past above might is an historical artifact.Historically, might was marked for past tense, just like could and would, for instance (cf. alsoStowell 2004). This defective past morphology can be re-activated by an embedded have, accordingto a suggestion by Hotze Rullmann and Lisa Matthewson.

– Support for such an analysis comes from the fact that modals without remnant past morphologysuch as may and can don’t allow for a past temporal perspective.

– A prediction is that an additional reading of (15) should be available for English, too, namely onewith epistemic possibility with both, past temporal perspective and past temporal orientation.This prediction seems to be borne out.

(66) Context:Mary is a school principal and at her school there is a policy that if there is even a possibilitythat a teacher has abused a student, the teacher will be fired. Five years ago, Mary fired one ofher teachers because he was accused of abusing a student. This morning, the accuser recantedthe accusation and conclusive proof was brought forward that that the accuser had lied andthe teacher was in fact innocent. Mary is now being interview by a report.Reporter: “How do you feel about the news today that the teacher you fired was in factinnocent?” – Mary: “Very upset. It is most unfortunate.” – Reporter: “So why did you fire himat the time, when you did not have conclusive proof that he was guilty?”(Lisa Matthewson and Hotze Rullmann)

(67) Because he might have abused that student. He might have been guilty.(past temporal, epistemic perspective, past temporal orientation)

– As the defective past tense requires an embedded perfect in order for it to be activated, we willhave to assume that in the case of a reading with future temporal orientation, there still is acovert prospective operator and the embedded have is interpreted as semantically vacuous.(I had not realized this before and I think this will require some refinement.)

(68) John might have won.

(69) hs, hi, tii

�w

0

�t

00 hti

past

hi, hhi, ti, tiit

00hi, ti

�t

0

mighths, hi, hhs, ti, tiii

w

0t

0hs, ti

�w

prosp

hi, hhi, ti, tiit

0have;

(past activation)

hi, ti

�t

pfv

hi, hhv, ti, tiit

hv, ti

�e

Johnwon w

e

14

Page 15: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

– A quick fix would be to allow for a higher, in this case the top most, �-operator to bind thetemporal variable introduced by have, an option allowed for under the current architecture andneeded e.g. for the temporal interpretation of relative clauses anyway (cf. also Hohaus and Beck2011; Kusumoto 1999, 2005; Percus 2000).

(70) hs, hi, tii

�w

0

�t

00 hti

past

hi, hhi, ti, tiit

00hi, ti

�t

0

mighths, hi, hhs, ti, tiii

w

0t

0hs, ti

�w

havehi, hhi, ti, tii

t

00hi, ti

�t

pfv

hi, hhv, ti, tiit

hv, ti

�e

Johnwon w

e

– In the resulting truth conditions, t and t

0 are not ordered with respect to each other, allowing fora future temporal orientation (as well as for past and present temporal orientations, of course).

(71) 9t [t < t

now

& 9w [Rw@,t(w) & 9t0 [t0 < t

now

& 9e [winw,e(John) & ⌧(e) ✓ t

0]]]]

5.2 Some Open Questions and Directions for Future Work

– Do we always have either a perfect or a covert prospective operator embedded?

– As discussed above, a future temporal orientation in the past with a past temporal perspec-tive we could also allow for by binding by a higher �-operator. There is a slight distinctionin truth conditions, however, that in this case we require that the event happened at somepoint before the utterance time. According to my intuition, this is what we want anyway.

– What does the architecture look like in the case of stative predicates?

– Stative predicates receive a present (i.e. simultaneous) temporal orientation rather than afuture temporal orientation. According to the analysis in Stechow (2009), stative predicatesdo not introduce an event argument. The denotation of the Verb Phrase will then be aproperty of times, the lower aspectual head will not be required.

Figure 17: Totem Field Storyboard: Feeding Fluffy

15

Page 16: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

(72) Fluffy might have been a dog.

Figure 18: Past Temporal, Epistemic Perspective and Present Temporal Orientation

– How can we tell in the case of a covert operator, specifically for temporal orientation, that alanguage has not incorporated this into the lexical entry of a particular modal?

– Only under the decompositional approach do we allow for binding of the temporal argumentintroduced by have or prosp by a higher �-operator. If these aspectual operators wereincorporated into the lexical entry of the modal, this option would be unavailable. Thisdifference should make predictions that might be testable.

– What is the semantic contribution of the subjunctive marking for counterfactuality in Germanand Dutch?

1. What is the LF architecture of tense and modality?

2. Do modals contribute to or restrict the temporal interpretation of a sentence?3. Does the grammar allow us to evaluate epistemic possibilities in the past?

1. I have specified a transparent LF architecture for times, worlds, and events that makes precisepredictions regarding the interaction of tense and modality.

2. No. The temporal perspective of the modal is provided by the temporal operator that binds thetime argument of the modals. The temporal orientation of the modal claim is provided by thetemporal and aspectual operators in the complement of the modal, which bind the time and eventarguments of the embedded predicated.

3. Yes. There is no interaction between the type of modality, which is contextually provided, andthe temporal interpretation of a sentence. Particularly, there are no grammatical, architecturalrestrictions on the temporal interpretation of epistemic possibility modals.

Acknowledgments

We are indebted to the various native speaker consultants who have contributed to this project. I wouldalso like to thank Sigrid Beck, Lisa Matthewson, and Hotze Rullmann as well as the other participantsof the ling530C seminar on modality in spring 2012 (University of British Columbia, Vancouver) forfeedback and discussion.

References

Alexiadou, Artemis, Monika Rathert, and Arnim von Stechow (2003). “Introduction: The Modules ofPerfect Constructions.” In: Perfect Explorations. Berlin: De Gruyter, pp. v–xxxvi.

Boogart, Ronny J. (2007). “The Past and Perfect of Epistemic Modals.” In: Recent Advances in the Syn-tax and Semantics of Tense, Aspect, and Modality. Ed. by Louis de Saussure, Jaques Moeschler,and Geneveva Puskas. Berlin: De Gruyter, pp. 47–70.

16

Page 17: Past Possibility Crosslinguistically - uni-tuebingen.de Possibility Crosslinguistically ... – Focus will mainly be on possibility modals. For ease of presentation, ... there are

Cinque, Guglielmo (1999). Adverbs and Functional Heads: A Cross-Linguistic Perspective. Oxford:Oxford University Press.

Condoravdi, Cleo (2002). “Temporal Interpretation of Modals: Modals for the Present and for thePast.” In: The Construction of Meaning. Ed. by David Beaver et al. Stanford: CSLI Publications,pp. 59–88.

Eide, Kristin M. (2003). “Modals and Tense.” In: Proceedings of Sinn und Bedeutung 7, pp. 120–135.Enç, Mürvet (1996). “Tense and Modality.” In: The Handbook of Contemporary Semantic Theory. Ed.

by Shalom Lappin. Oxford: Blackwell, pp. 345–358.Fintel, Kai von and Anthony S. Gillies (2008). “CIA Leaks.” In: Philosophical Review 117.1, pp. 77–98.Groenendijk, Jeroen and Martin Stokhof (1975). “Modality and Conversational Information.” In: The-

oretical Linguistics 2, pp. 61–112.Hacquard, Valentine (2006). “Aspects of Modality.” PhD thesis. Cambridge: Massachusetts Institute

of Technology.Hohaus, Vera and Sigrid Beck (2011). “Lecture Notes on Tense and Modality.” Eberhard Karls Univer-

sität Tübingen. Feb. 2011.Kratzer, Angelika (1981). “The Notional Category of Modality.” In: Words, Worlds, and Contexts: New

Approaches in Word Semantics. Ed. by Hans-Jürgen Eikmeyer and Hannes Rieser. Berlin: DeGruyter, pp. 38–74.(1991). “Modality.” In: Semantik: Ein internationales Handbuch der zeitgenössischen Forschung.Ed. by Arnim von Stechow and Michael Herwig. Berlin: De Gruyter, pp. 639–650.

Kusumoto, Kiyomi (1999). “Tense in Embedded Context.” PhD thesis. Amherst: University of Mas-sachusetts.(2005). “On the Quantification over Times in Natural Language.” In: Natural Language Seman-tics 13.4, pp. 317–357.

Matthewson, Lisa (2004). “On the Methodology of Semantic Fieldwork.” In: International Journal ofAmerican Linguistics 70.4, pp. 369–451.(2011). “Methods in Crosslinguistic Formal Semantics.” In: Semantics: An International Hand-book of Natural Language Meaning. Ed. by Claudia Maienborn, Klaus von Heusinger, and PaulPortner. Vol. 1. Berlin: De Gruyter, pp. 268–284.(to appear). “Gitksan Modals.” In: International Journal of American Linguistics.

Percus, Orin (2000). “Constraints on Some Other Variables in Syntax.” In: Natural Language Semantics8, pp. 173–229.

Peterson, Tyler (2010). “Epistemic Modality and Evidentiality in Gitksan at the Semantics/Pragmatics-Interface.” PhD thesis. Vancouver: University of British Columbia.

Stechow, Arnim von (1999). “Eine erweiterete Extended Now-Theorie für Perfekt und Futur.” In:Zeitschrift für Literaturwissenschaft und Linguistik 113, pp. 86–118.(2009). “Tenses in Compositional Semantics.” In: The Expression of Time. Ed. by Wolfgang

Klein and Ping Li. Berlin: De Gruyter, pp. 129–166.Stechow, Arnim von and Sigrid Beck (to appear). “Events, Times, and Worlds: An LF Architecture.” In:

Situationsargumente im Nominalbereich. Ed. by Christian Fortmann et al. Tübingen: NiemeyerVerlag.

Stowell, Timothy Osborne (2004). “Tense and Modals.” In: The Syntax of Time. Ed. by JacquelineGuéron and Jacqueline Lecarme. Cambridge: The MIT Press, pp. 621–636.

17