The 15th Japanese/Korean Linguistics Conference
(The Univ. of Wisconsin-Madison, Oct.7-9/05)
Scrambling, resumption, and scope of Neg
Oct./08/05 Kiyoko KATAOKA
(Nihon University, College of Economics)
Outline
1. Introduction 1
1.1. Scope ambiguity 1
1.2. Absence of scope ambiguity 2
1.3. A/A'-properties of 'scrambled' object 2
1.4. Goal 3
2. Account 1: A-movement analysis for A-scrambling (Miyagawa 2001 and others) 3
2.1. Absence of Neg>QP for A-ScramObj 3
2.2. Problem for A-movement analysis: 'resumption' 4
3. Account 2: Da derivational account of A/A'-properties of (Saito 2003) 4
4. Account 3: Bbase-generation analysis (Ueyama 1998) 5
4.1. Absence of Neg>QP for A-ScramObj 5
4.2. 'Resumption' under the base-generation analysis 6
5. Further A supporting argument for the base-generation analysis of A-ScramObj 6
5.1. A-ScramObj and 'resumption' 6
5.2. Absence of Neg>QP for A-ScramObj 7
6. Conclusion 7
References 7
1. Introduction
1.1. Scope ambiguity
♦ The scope interpretation between Neg and QP is ambiguous in Japanese (e.g., Kuno 1980).[1]
(1) a. [QP itutuizyoo-no ginkoo ]-ga Toyota-ni yuusisi-nak-atta (koto)
5:or:more-GEN bank-NOM Toyota-DAT finace-Neg-Past (Comp)
'Five or more banks did not finance Toyota.' 5-or-moreNeg, Neg5-or-more
b. UFJ-ginkoo-ga [QP itutuizyoo-no kigyoo ]-ni yuusisi-nak-atta (koto)
UFJ Bank-NOM 5:or:more-GEN company-DAT finance-Neg-Past (Comp)
'UFJ Bank did not finance five or more companies.' 5-or-moreNeg, Neg5-or-more
As in (2), tThe ambiguity in (1) continues to obtain in their 'scrambled' counterparts in (2).
(2) a. Toyota-ni [QP itutuizyoo-no ginkoo ]-ga yuusisi-nak-atta (koto)
Toyota-DAT 5:or:more-GEN bank-NOM finace-Neg-Past (Comp)
5-or-moreNeg, Neg5-or-more
b. [QP itutuizyoo-no kigyoo ]-ni UFJ-ginkoo-ga yuusisi-nak-atta (koto)
5:or:more-GEN company-DAT UFJ Bank-NOM finance-Neg-Past (Comp)
5-or-moreNeg, Neg5-or-more
1.2. Absence of scope ambiguity
♦ If the bound variable anaphora (BVA) interpretation is forced to obtain between [the object QP1 <= This is in fact ScramObj , but you introduce the term later. In the class on Thursday, it was suggested that yu may want to first introduce the term ScramObj, and use it in place of "object QP1" and " the 'scrambled' object (henceforth ScramObj)" in this sentence.] and the dependent term soko 'it' contained in inside the subject NP, however, the 'scrambled' object (henceforth ScramObj) cannot be in the scope of Neg. (Cf. (2b).)
(3) [Itutuizyoo-no kigyoo]QP1-ni [soko-no torihiki-ginkoo]-ga yuusisi-nak-atta.
5:or:more-GEN-company-DAT it-GEN-dealing:bank-NOM finance-Neg-Past
'To five or more companies, its own bank did not finance.'
with BVA, *Neg > QP1
1.3. A/A'-properties of 'scrambled' object
(4) Scrambling construction
a. A'-properties: e.g., reconstruction effects ((5a))
b. A-properties: e.g., the absence of weak crossover effects ((5b))
(Saito 1992, Ueyama 1998, among others).
(5) a. Sokoi-no torihikisaki-ni [itutuizyoo-no kigyoo]QP1-ga kitaidoori tousisi-ta.
it-GEN client-DAT 5:or:more-GEN company-NOM as expected invest-Past
'To itsi client, [each of five or more companies]i invested as ecpected.'
ok BVA
b. [Itutuizyoo-no kigyoo]QP1-ni sokoi-no torihikisaki-ga kitaidoori tousisi-ta.
5:or:more-GEN company-DAT it-GEN client-NOM as expected invest-Past
'To [each of five or more companies]i, itsi client invested as expected.'
ok BVA
(6) In order for the BVA interpretation to obtain between a QP α and a dependent term b, α must c-command b at pre-QR position (Reinhart 1983, Ueyama 1998, Hoji 2003).
♦ With the BVA interpretation in question, the subject must c-command the object in (5a), and the object must c-command the subject at LF in (5b), at their pre-QR positions of LF.
♦ For 'scrambled' sentences, The two kinds of distinct c-command relations must obtain at LF between the subject and the object are available to the 'scrambled' sentences. The former corresponds to A'-properties and the latter to the A-properties.
(7) Scrambling construction: NP-ACC/DAT NP-NOM V
a. LF1: [ NP-NOM [ NP-ACC/DAT V ] ] → A'-properties / A'-scrambling → (5a)
b. LF2: [ NP-ACC/DAT [ NP-NOM V ] ] → A-properties / A-scrambling → (5b)
♦ The absence of the Neg>QP reading in example (3) shows that ScramObj the scrambled object with A-properties (henceforth A-ScramObj) cannot be in the scope of Neg.
(3) [Itutuizyoo-no kigyoo]QP1-ni [soko-no torihiki-ginkoo]-ga yuusisi-nak-atta.
5:or:more-GEN-company-DAT it-GEN-dealing:bank-NOM finance-Neg-Past
'To five or more companies, its own bank did not finance.' with BVA, *Neg > QP1
1.4. Goal
(8) To [argue => account?] for the absence of the Neg>NP reading for the A-ScramObj NP.
(9) To argue against (a) and (b), and argue in support of (c).
a. Account 1: an A-movement analysis of 'A-scrambling' (Miyagawa 2001 among others)
b. Account 2: a derivational account of A/A'-properties of 'scrambling' (Saito 2003)
c. Account 3: a base-generation analysis of 'A-scrambling' (Ueyama 1998)
♦ I will assume, without further discussion:,
(10) [Assumption 1: The scope of α is its c-command domain at LF, and Neg is in the VP (or vP)-adjoined position at LF. <= You may want to break this into two parts. If you do that, it might be easier to provide some basic references for each of them in the parentheses after each assumption. You don't have to read them in your presentation; the audience can see the references in the handout.]
2. Account 1: A-movement analysis for of A-scrambling (Miyagawa 2001 and others)
(9) a. Account 1: an A-movement analysis of 'A-scrambling' (Miyagawa 2001 among others)
2.1. Absence of Neg>QP for A-ScramObj
♦ The A-moved NP undergoes 'quantifier-lowering' (May 1977).
(11a) allows the reading in (11b).
(11) a. Someone is likely to address the rally.
b. It is likely that someone will address the rally.
♦ LF-lowering of the A-moved NP is blocked by a negative predicate (Lasnik & Saito 1991).
(12a) does not allow the reading in (12b).
(12) a. Someone is unlikely to address the rally.
b. It is unlikely that someone will address the rally.
(13) Neg blocks 'quantifier-lowering' for the A-moved NP.
(14) The A-ScramObj is moved to its surface position from its theta-position by A-movement (Miyagawa 2001, among others).
♦ The absence of the Neg>QP reading for the A-ScramObj as in (3) can be accounted for, by attributing it to blocking by Neg of 'quantifier-lowering' for the A-moved NP.[2]
2.2. Problem for A-movement analysis: 'resumption'
(15) A-movement analysis: The A-properties of ScramObj come from A-movement.
♦ What appears to be 'resumption' is compatible with A-properties (cf. (5b)), but it is not compatible with A'-properties (cf. (5a)), as discussed in Hayashishita 1997, Hoji & Ueyama 1998 and Hoji 2003.
(16) [Itutuizyoo-no kigyoo]QP1-ni sokoi-no torihikisaki-ga soko-ni tousisi-ta.
5:or:more-GEN company-DAT it-GEN client-NOM it-DAT invest-Past
'To five or more companiesi, itsi client invested to it.' (with BVA)
(17) *Sokoi-no torihikisaki-ni [itutuizyoo-no kigyoo]i-ga soko-ni tousisi-ta.
it-GEN client-DAT 5:or:more-GEN bank-NOM it-DAT invest-Past
♦ Crucially, the ScramObj in (16), which is A-ScramObj, cannot be in the scope of Neg.
(18) [Itutuizyoo-no kigyoo]QP1-ni sokoi-no torihikisaki-ga soko-ni tousisi-nak-atta.
5:or:more-GEN company-DAT it-GEN client-NOM it-DAT invest-Neg-Past
with BVA(QP1, soko), * Neg > QP1, ok QP1 > Neg
I think this is anti-climactic. I.e., it would be more effective if you present your argument more in a step-by-step way. E.g.,
--The A-movement-based account of the absence of the Neg>QP reading would say, as you have mentioned above, that the lowering of QP1 is bloced by Neg, given that A-ScramObj has undergone A-movement.
--But there is an independent observation that 'resumption' is compatible with A-properties but not with A'-properties.
--Hence, the A-movement-based analysis the A-properties in question is suspect, given that A-movement does not allow resumption.
--Now the climax: In fact, even without BVA, the Neg>QP reading is not available as long as we have 'resumption' as in (xx).
(xx) an example without BVA
--Less climactic: If we consider an example with BVA and resumption, it does not allow the Neg>QP reading, as we now expect. Then give (18).
(19) 'Resumption' seems compatible with A-properties but not with A'-properties.[3] <= To make this point, do you need (18)? I.e., it is not clear whether you need (18) in arguing against the A-movement analysis. Do you? I think you should either eliminate (18) here and put is in a later section where you would argue for the base-generation analysis, or present it as the less climatic 'element' as indicated/suggested above.
♦ The observation is inconsistent with the general assumption that A-movement does not allow resumption.
3. Account 2: a derivational account of A/A'-properties of (Saito 2003)
(9) b. Account 2: a derivational account of A/A'-properties of scrambling (Saito 2003)
(20) Saito's (2003) derivational analysis for A/A'-properties of ScramObj:
a. The relevant formal relation underlying the A/A'-properties is established or checked derivationally.
b. Every ScramObj must be 'reconstructed' at LF.
♦ If a scope relation between α and b were is assumed to obtain based on their c-command relation at some stage of derivation, every ScramObj could be in the scope of Neg, since even an A-ScramObj would certainly be in the c-commanded domain of Neg while being it is in its theta-position.
♦ The absence of the Neg>QP reading for the A-ScramObj QP would be problematic.
4. Account 3: base-generation analysis (Ueyama 1998)
(9) c. Account 3: a base-generation analysis of 'A-scrambling' (Ueyama 1998)
(21) Ueyama's (1998) anayisis
a. The A-ScramObj is base-generated in its surface-position c-commanding the subject.
b. Null-operator movement, like the one in the English tough construction, is assumed in order to make a syntactic relation between the theta-position and the A-ScramObj.
♦ [The observations above <= Make it transparent which observations, by providing a pointer of some sort.] can be accounted for without any problem.
4.1. Absence of Neg>QP for A-ScramObj
♦ The absence of the Neg>QP reading for the A-ScramObj as in (3) is as predicted.
(3) [Itutuizyoo-no kigyoo]QP1-ni [soko-no torihiki-ginkoo]-ga yuusisi-nak-atta.
5:or:more-GEN-company-DAT it-GEN-dealing:bank-NOM finance-Neg-Past
'To five or more companies, its own bank did not finance.' with BVA, *Neg > QP1
♦ Given (6), in order for the BVA reading to obtain in (3), the object QP should be base-generated in the position c-commanding the subject NP, as illustrated in (22).
(6) In order for the BVA interpretation to obtain between a QP α and a dependent term b, α must c-command b at pre-QR position (Reinhart 1983, Ueyama 1998, Hoji 2003).
(22) Pre-QR of LF: [ [ ... ]QP1-DAT [VP [ ... soko ... ]-NOM V ] ]
♦ The ScramObj QP, therefore, is never in the c-command domain of Neg throughout its derivation, provided (23).,
(23) a. that itsThe base-generated position for A-ScramObj is outside the c-command domain of Neg (i.e., VP), as extensively argued in Kataoka 2004 and to appear on independent grounds, and.
b. that Mmovement can never be downward.
(24) Structure before QR of (3): [ [ ... ]QP1-DAT [NegP [Neg' [VP [ ... soko ... ]-NOM V] [Neg -nai] ] ] ]
(Cf.) Sokoi-no torihikisaki-ni [itutuizyoo-no kigyoo]QP1-ga tousisi-nak-atta.
it-GEN client-DAT 5:or:more-GEN company-NOM invest-Neg-Past
'To itsi client, [each of five or more companies]i did not invest.'
ok {BVA & Neg > QP1}
4.2. 'Resumption' under the base-generation analysis
♦ The possibility of 'resumption' is precisely what is expected in the case of A-scrambling.
(25) 'Resumption' is an overt counterpart of an empty argument which occupies the theta-position, and functions like a null operator in the English tough construction in order to make a syntactic relation between the theta-position and the A-ScramObj. (Hoji & Ueyama 1998, Hoji 2003)
♦ The A-ScramObj in (18) with 'resumption' cannot be in the c-command domain of Neg at LF, either, since the ScramObj is base-generated in the position outside the domain of Neg.
As suggested above, you may want to include (18) and other related materials here.
(18) [Itutuizyoo-no kigyoo]QP1-ni sokoi-no torihikisaki-ga soko-ni tousisi-nak-atta.
5:or:more-GEN company-DAT it-GEN client-NOM it-DAT invest-Neg-Past
with BVA(QP1, soko), * Neg > QP1, ok QP1 > Neg
5. A Further support ing argument for the base-generation analysis of A-ScramObj
♦ QP-scope interpretations involving resumption and Neg show parallelism with the BVA cases.
5.1. A-ScramObj and 'resumption'
♦ The scope relation is ambiguous in the scrambling construction.
(26) [Mittuizyoo-no kigyoo]QP2-ni [itutuizyoo-no ginkoo]QP1-ga kitaidoori yuusisi-ta.
3:or:more-GEN company-DAT 5:or:more-GEN bank-NOM as expected finance-Past
'To three or more companies, five or more banks financed as expected.'
DR(QP1>QP2), DR(QP2>QP1)
(27) In order for the the DR(QP1>QP2) reading to obtain, the QP1 must c-command the QP2 at LF. (Reinhart 1983, Hayashishita 2004)
♦ Given (27), the subject QP1 must c-command the object QP2 to take wide-scope, while the object QP2 must c-command the subject QP1 to take wide-scope.
(28) a. The establishment of DR(QP1>QP2) in (26) → A'-property
b. The establishment of DR(QP2>QP1) in (26) → A-property <= In your discussion of BVA, you had plural "properties" …
♦ Resumption is possible only when the ScramObj QP takes scope over the subject QP, i.e., when it shows A-property.
(29) a. [Mittuizyoo-no kigyoo]QP2-ni [itutuizyoo-no ginkoo]QP1-ga kitaidoori soko-ni yuusisi-ta.
3:or:more-GEN company-DAT 5:or:more-GEN bank-NOM as expected it-DAT finance-Past
'To three or more companies, five or more banks financed to it as expected.'
ok { 'resumption' soko-ni & DR(QP2>QP1)}
b. [Mittuizyoo-no kigyoo]QP2-ni [itutuizyoo-no ginkoo]QP1-ga kitaidoori soko-ni yuusisi-ta.
3:or:more-GEN company-DAT 5:or:more-GEN bank-NOM as expected it-DAT finance-Past
* { 'resumption' soko-ni & DR(QP1>QP2)}
5.2. Absence of Neg>QP for A-ScramObj
♦ With the DR reading for the ScramObj QP2 over the QP1, which shows the ScramObj QP2 is an A-ScramObj, Neg>QP2 is impossible.
(30) [Mittuizyoo-no kigyoo]QP2-ni [itutuizyoo-no ginkoo]QP1-ga yuusisi-nak-atta.
3:or:more-GEN company-DAT 5:or:more-GEN bank-NOM finance-Neg-Past
'To three or more companies, five or more banks did not financed.'
*{DR(QP2>QP1) & Neg>QP2}, but ok{DR(QP1>QP2) & Neg>QP2}
♦ The ScramObj QP involving resumption cannot be in the scope of Neg as well.
(31) [Mittuizyoo-no kigyoo]QP2-ni [itutuizyoo-no ginkoo]QP1-ga soko-ni yuusisi-nak-atta.
3:or:more-GEN company-DAT 5:or:more-GEN bank-NOM it-DAT finance-Neg-Past
[ok { 'resumption' soko-ni & DR(QP2>QP1)}, but then *Neg>QP2 <= What is intended by this part is not transparent, anyway.]
As in the case of the discussion re. (18), (31) is in a way less climatic than (i), which should be compared with (ii).
(i) [Mittuizyoo-no kigyoo]QP2-ni City Bank-ga soko-ni yuusisi-nak-atta.