Proper Binding Condition

in Koizumi 2000, Takano 2002 & Kawazoe 2005

1.Koizumi 2000

1.1Koizumi's suggestions

(K98) Proper Binding Condition:

Traces must be bound. (Fiengo, 1977; May, 1977)

Koizumi's suggestions:

(1)The PBC does not apply at S-structure.

(2)What has been regarded as a violation of the PBC should be taken care of by (i) and (ii) below:

(i)Koizumi 2000:266:(100)

A Restatement of Müller's generalization:

If dominates a trace of which has been created by a certaintype of movement, cannot undergo the same type of movementyielding the structure in which the trace of in is unbound.

(adapted from Müller (1993, 1994) with modification by M.K.)

(ii)A-bar traces (variables) must be bound at LF.

1.2.Previous Study

Saito (1989) argues that the PBC applies at not only at LF but also at S-structure. (K99c) is considered as evidence for that.

(K99)a. John-ga [Bill-ga sono mura-ni sundeiru to]omotteiru (koto).

John-NOM [Bill-NOM that village-in reside COMP]think

‘John thinks that Bill lives in that village.’

b.Sono mura-niiJohn-ga [Bill-ga tisundeiru to]omotteiru (koto).

that village-ini John-NOM[Bill-NOM ti reside COMP]think

(I)

c.* [Bill-ga tisundeiru to]j sono mura-nii John-gatjomotteiru (koto).

[Bill-NOM ti reside COMP]j that village-ini John-NOMtj think

(II)

(cf. Saito, 1985, 1989)

(3)Assumptions

a.Long-distance scrambling is an instance of A'-movement.

b.A'-movement can freely be undone.

If the PBC applied at LF:

Under the assumptions in (3), the moved element in (K99c) would be undone at LF, and there would be no unbound trace which would violate the PBC. Hence, the unacceptable status of the sentence in (K99c) would remain unexplained.

If the PBC applied at SS:

The PBC applies to the structure in (K99c) in that the trace of sono mura-ni is unbound. The unacceptable status of (K99c) could successfully be accounted for.

Thus, the case in (K99c) has been considered as evidence that the PBC applies at SS as well as LF.

1.3.Müller's generalization

1.3.1Definition and Application

There have been a number of attempts to eliminate S-structure application of the PBC. Most of the proposals is made to deriveMüller's generalization.

(K100)A Restatement of Müller’s generalization

If dominates a trace of which has been created by a certaintype of movement, cannot undergo the same type of movementyielding the structure in which the trace of in is unbound.

(adapted from Müller (1993, 1994) with modification by M.K.)

(3)Assumptions

a.Long-distance scrambling is an instance of A'-movement.

b.A'-movement can freely be undone.

Given (K100) and (3), (K99c) would have the two possible derivations, both of which would result in ungrammaticality of the sentence.

(K99)a. John-ga [Bill-ga sono mura-ni sundeiru to]omotteiru (koto).

John-NOM [Bill-NOM that village-in reside COMP]think

‘John thinks that Bill lives in that village.’

b.Sono mura-niiJohn-ga [Bill-ga tisundeiru to]omotteiru (koto).

that village-ini John-NOM[Bill-NOM ti reside COMP]think

(I)

c.* [Bill-ga tisundeiru to]j sono mura-nii John-gatjomotteiru (koto).

[Bill-NOM ti reside COMP]j that village-ini John-NOMtj think

(II)

(cf. Saito, 1985, 1989)

(K101) Two possible derivations of (K99c):

(I) / (II)
Derivation I / A-bar / A / *PBC violation at LF
Derivation II / A-bar / A-bar / *violation of Müller’s generalization

As shown in (K101), the sentence in (K99c) which have been considered as evidence for S-structure application of the PBC would be accounted for by LF application of the PBC together with Müller’s generalization.

1.3.2Müller’s generalization vs. S-structure application of PBC

Examples which would be derived in the steps illustrated in (K102) would be predicted to be ungrammatical by S-structure application of the PBC but grammatical by Müller’s generalization.

(K102)a....X...[...Y...]...

b.Yi...X...[...ti...]...

A-scrambling

c.[...ti...]j ...Yi...X...tj...

A'-scrambling

Let us look at an example which corresponds to the derivation illustrated in (K102c):

(K103)a.(?)[[Tom-ga Mary-ni ringo-o 2-tu] to [Bob-gaMary-ni banana-o

[[Tom-NOM Mary-toapple-ACC 2-CL] and [Bob-NOMMary-to banana-AC

3-bon]] ageta (koto).

C 3-CL]] gave]

b.Mary-nii [[Tom-ga ti ringo-o 2-tu] to [Bob-ga tibanana-o 3-bon]]

Mary-toi [[Tom-NOM ti apple-ACC 2-CL] and [Bob-NOM tibanana-ACC 3-CL]]

(I) “across-the-board” scrambling

ageta (koto).

gave

c.[[Tom-ga ti ringo-o 2-tu] to [Bob-ga ti banana-o 3-bon]]j Mary-nii

[[Tom-NOM tiapple-ACC 2-CL] and [Bob-NOM tibanana-ACC 3-CL]]j Mary-toi

(II)

tj ageta (koto).

tjgave

There are four logically possible combinations of movements involved in (I) and (II):

(K104)

(I) / (II) / Koizumi analysis:
PBC at LF & Müller’s / Saito's analysis:
PBC at SS & LF
Case I / A-bar / A-bar / *Müller’s generalization / *PBC at SS
Case II / A / A / *Müller’s generalization / *PBC at LF
Case III / A-bar / A / *PBC violation in LF / *PBC at LF
Case IV / A / A-bar / ✓ / *PBC at SS

Koizumi's analysis can but Saito's analysis cannot account for the acceptable status of (K103c).

Koizumi states:

"The grammaticality of ([K] 103c) strongly suggests that the PBC as an S-structure requirement does not exist."

The same point can be illustrated for the following example.

(K107c)?[[Tom-ga ti ringo-o 2-tu] to [Bob-ga tibanana-o 3-bon]]j

[[Tom-NOM ti apple-ACC 2-CL] and [Bob-NOM ti banana-ACC 3-CL]]j

Mary-nii John-ga [tj ageta to]omotteiru (koto).

Mary-toi John-NOM [tj gave that]believe

For (K107c), there are two possible derivations which does not violate Müller’s generalization.

Possible derivation 1 for (K107c)

(K108)a.?John-ga [[[Tom-ga Mary-ni ringo-o 2-tu] to[Bob-ga Mary-ni

John-NOM [[[Tom-NOM Mary-to apple-ACC 2-CL] and[Bob-NOM Mary-to

banana-o 3-bon]] ageta to]omotteiru (koto).

banana-ACC 3-CL]] gave that]believe

b.John-ga [Mary-nii [[Tom-ga ti ringo-o 2-tu] to[Bob-ga ti banana-o

John-NOM [Mary-toi [[Tom-NOM ti apple-ACC 2-CL] and[Bob-NOM ti banana-ACC

(I) ATB A-scrambling

3-bon]] ageta to]omotteiru (koto).

3-CL]] gave that] believe

c.(?)Mary-nii John-ga [ti[[Tom-ga ti ringo-o2-tu] to[Bob-ga ti

Mary-toi John-NOM [ti[[Tom-NOM ti apple-ACC2-CL] and [Bob-NOM ti

(II) A-bar scrambling

banana-o 3-bon]] agetato] omotteiru (koto).

banana-ACC 3-CL]] gavethat] believe

d.? [[Tom-ga tiringo-o 2-tu] to [Bob-ga tibanana-o 3-bon]]j

[[Tom-NOM ti apple-ACC 2-CL] and [Bob-NOM tibanana-ACC 3-CL]]j

Mary-nii John-ga [titj agetato] omotteiru (koto).

Mary-toi John-NOM [titj gave that] believe

(III) A-bar scrambling

The movements in (I) and (III) are not of the same kind. Hence, the movement which creates the shaded trace, ti is not a violation of Müller’s generalization.

Possible derivation 2 for (K107c)

(K109)a.?John-ga [[[Tom-ga Mary-ni ringo-o 2-tu] to[Bob-ga Mary-ni

John-NOM [[[Tom-NOM Mary-to apple-ACC 2-CL] and[Bob-NOM Mary-to

banana-o 3-bon]] ageta to]omotteiru (koto).

banana-ACC 3-CL]] gave that]believe

b.(?)[[Tom-ga Mary-ni ringo-o 2-tu] to [Bob-gaMary-ni banana-o

[[Tom-NOM Mary-to apple-ACC 2-CL] and [Bob-NOMMary-to banana-ACC

3-bon]]j John-ga [tj ageta to]omotteiru (koto).

3-CL]]j John-NOM [tj gave that]believe

(I) A-scrambling

c.?Mary-nii [[Tom-ga ti ringo-o 2-tu] to [Bob-ga ti banana-o

Mary-toi [[Tom-NOM ti apple-ACC 2-CL] and[Bob-NOM ti banana-ACC

(II) ATB A-scrambling

3-bon]]j John-ga [tj agetato] omotteiru (koto).

3-CL]]j John-NOM [tj gavethat] believe

d.?[[Tom-ga tiringo-o 2-tu] to [Bob-ga tibanana-o 3-bon]]j

[[Tom-NOM tiapple-ACC 2-CL] and [Bob-NOM tibanana-ACC 3-CL]]j

Mary-nii tjJohn-ga [tj agetato] omotteiru (koto).

Mary-toi tjJohn-NOM [tj gavethat] believe

(III) A-bar scrambling

1.3.3A-movement and head-movement

Based on the acceptable status of the examples in (K111), Koizumi concludes that it is likely that traces of A-moved or head-moved elements are not subject to the PBC.

(K111)a.Mary-ga agetav no-wa [John-ni ringo-o 3-tu tv] da.

Mary-NOM gavev NL-TOP [John-to apple-ACC 3-CL tv] be

Lit. ‘It is [three apples to John] that Mary gave.’

b.[How (ti) likely (ti) to ti win] is Johni?

c.what Johni is [is (ti) likely (ti) to ti win]

1.4Summary

---It seems likely that traces of A-moved or head-moved elements are immune to the PBC.

---The PBC should be restated as a sub-case of the requirement that variables must be bound at LF;

(i) A-bar traces (variables) must be bound at LF.

----PBC should be eliminated from grammar as an independent syntactic condition.

2.Takano 2002

2.1Oblique Movement Approach

(T89)[John-ga ringo-o mittu] to [Mary-ga and banana-o nihon] katta.

=(T76)[John-Nom apple-Acc three.C [Mary-Nom L ban ana-Acc two.CL bought

‘John bought three apples and Mary two bananas.’

Suppose UG has a way of conjoining derived constituents.

Assumptions:

(a) Every syntactic structure is built up by successive application of the operations Merge and Move.

(b) Merge and Move can be carried out by either substitution or adjunction (Chomsky (1995)).

Under the assumptions, coordination should be able to apply in the middle of structure building.

(T90’) banana-o nihon is scrambled to TP and Mary-ga is adjoined to banana-o

nihon by oblique movement.

[TP [Y [x Mary-ga] banana-o nihon] [TP tx tY V]]

(T91’) John-ga and ringo-o mittu are put together independently of (T90’), and it adjoins to a coordinator as in (T91’)

a. [Z [W John-ga] ringo-o mittu]

b.[[Z [W John-ga] ringo-o mittu] &]

(T92’) the constituent in (T91’) adjoins to (T90’).

[TP [Y[[Z [W John-ga] ringo-o mittu] &] [Y [X Mary-ga] banana-o

nihon]] [TP tx tY katta]]

Let us see how the sentence which is phonetically identical to the sentence in (K103) can be derived under the analysis pursued in Takano 2002.

(K103c)[[Tom-ga ti ringo-o 2-tu] to [Bob-ga ti banana-o 3-bon]]j Mary-nii

[[Tom-NOM tiapple-ACC 2-CL] and [Bob-NOM tibanana-ACC 3-CL]]j Mary-toi

(II)

tj ageta (koto).

tjgave

(4)a.banana-o 3-bonis scrambled to TP and Bob-ga is adjoined to banana-o 3-bon by oblique movement.

[TP [Y[XBob-ga] banana-o 3-bon] [TP tXMary-nitYageta]]

b.Tom-ga and ringo-o 2-tu are put together independently of the structure in (4a), and it adjoins to a coordinator as follows;

(i)[Z [WTom-ga] ringo-o 2-tu]

(ii)[[Z [WTom-ga] ringo-o 2-tu] &]

c.the constituent in (4b-ii) adjoins to (4a)

[TP [Y[[Z [WTom-ga] ringo-o 2-tu] &] [Y [XBob-ga] banana-o

3-bon]] [TP tx tYMary-ni ageta]]

There is no unbound trace in the coordination structure. Hence, the PBC is irrelevant here.

(K107c)?[[Tom-ga ti ringo-o 2-tu] to [Bob-ga tibanana-o 3-bon]]j

[[Tom-NOM ti apple-ACC 2-CL] and [Bob-NOM ti banana-ACC 3-CL]]j

Mary-nii John-ga [tj ageta to]omotteiru (koto).

Mary-toi John-NOM [tj gave that]believe

(5)a.banana-o 3-bonis scrambled to TP2 and Bob-ga is adjoined to banana-o 3-bon by oblique movement.

[TP1John-ga [TP1[Y[XBob-ga] banana-o 3-bon] [TP2 txMary-ni tYageta]]]-to omotteiru]

b.Tom-ga and ringo-o 3-tu are put together independently of the structure in (5a), and it adjoins to a coordinator as follows;

(i)[Z [WTom-ga] ringo-o 2-tu]

(ii)[[Z [WTom-ga] ringo-o 2-tu] &]

c.the constituent in (5b-ii) adjoins to (5a)

[TP1 John-ga[VP [V'[TP2 [Y[[Z [WTom-ga] ringo-o 2-tu] &] [Y [X Bob-ga] banana-o 3-bon]] [TP2 txMary-ni tYageta]]-to omotteiru]

d.Mary-ni is scrambled to the matrix clause.

[TP1 Mary-niZ[TP1 John-ga[TP2 [Y[[Z [WTom-ga] ringo-o 2-tu] &] [Y [XBob-ga] banana-o 3-bon]] [TP2 txtz tYageta]]-to

omotteiru]A'-movement

e.The coordinated structure (=Y) is scrambled to the sentence-initial position.

[TP1[Y[[Z [WTom-ga] ringo-o 2-tu] &] [Y [XBob-ga] banana-o 3-bon]] u[TP1 Mary-niZ [TP1 John-ga [TP2tu[TP2 txtz tY ageta]]-to omotteiru] A'-movement

No trace is inside the coordination structure, hence, no PBC violation.

3.Kawazoe 2005

Under Kawazoe (2005)'s analysis, (K1) is analyzed as having the structure in (K37).

(K1)[Taroo-ga ringo (o mittu)] to [Ziroo-ga mikan-o (futatu) ] kau

(K37)[VP[pP [pPTaroo-ga [p' ringo (o mittu) p ] to [pP Ziroo-ga [p'mikan-o (futatu) p]] kau]

Let us see how the sentenceswhich are phonetically identical to the sentences in (K103) and (K107c), repeated below, can be derived under Kawazoe (2005)'s analysis.

(K103c)[[Tom-ga ti ringo-o 2-tu] to [Bob-ga ti banana-o 3-bon]]j Mary-niitj

[[Tom-NOM tiapple-ACC 2-CL] and [Bob-NOM tibanana-ACC 3-CL]]j Mary-toitj

ageta (koto).(II)

gave

(6)Possibility 1

[VP[pP [pPTaroo-ga [p' ringo-o 2-tu [p] ] to [pP Bob-ga [p'banana-o 3-bon [p]][V' Mary-ni [V ageta ]]] (koto)

----we have to assume that NP-o can be base-generated in the place that c-commands NP-ni. In that case, "格配列照合操作" do not seem to work.

格配列照合操作

(K41)辞書における動詞の格配列の指定

壊す: -[を、が]

壊れる: -[が]

(K42)辞書における格助詞の格配列の指定

が:+[が]

を:+[を]

(K43)格配列照合原理:

格配列中の格素性はすべて照合されなければならない。

There are three operations: ｢照合」「浸透」「結合」

(K44)a.格配列の照合

+の格配列と-の格配列の間で起こる。

複数の格素性を含んでいる場合、最初のものだけが可視的。

(K44)

(成功例1)-[case1]

+[case2] -[case2, case1]

(失敗例1)*

+[case2]-[case1, case2]

[VP[pP [pPTaroo-ga [p' ringo-o 2-tu [p] ] to [pP Bob-ga [p'banana-o 3-bon [p]][V' Mary-ni [V ageta ]]] (koto)

VP

pP

pPtopP

Taro-ga p Bob-ga p V'

ringo-o futatu p banana-o 3-bon p Mary-niageta.

+[ni] -[o, ni, ga]

失敗

(7)Possibility 2pP can be created by movement.

a.ringo-o 2-tubanana-o 3-bonundergo movement.

[VP[pP [pPTaroo-ga[p'[ringo-o 2-tu]i [p] ] to [pP Bob-ga [p'[banana-o 3-bon]i [p]][ Mary-ni [ ti [V ageta ]]] (koto)

b.NP-o is base-generated as Deep DL and NP-ga is scrambled to the sentence-initial position.

[XP [pP [pPTaroo-gaj [p' [ringo-o 2-tu] [p] ]to [pP Bob-ga [p'[banana-o 3-bon] [p]] [VPtj[ Mary-ni [ [V ageta ]]] (koto)

----we have to assume that pP can be generated not only at base-structure but also by movement, which would make the proposal similar to Takano 2002.

---Generally, movement crossing over Deep DL is only marginally acceptable. If the derivation in (7b) is correct, the sentence would be unacceptable due to NP-ga moving across Deep DL.

Conclusion:

None of the possibilities above seems plausible. We cannot conclude how PBC would apply or not apply to coordination sentences under Kawazoe's analysis.

For the same reason, I do not understand how we can derive the sentence which is phonetically identical to that in (K107c)

(K107c)?[[Tom-ga ti ringo-o 2-tu] to [Bob-ga tibanana-o 3-bon]]j

[[Tom-NOM ti apple-ACC 2-CL] and [Bob-NOM ti banana-ACC 3-CL]]j

Mary-nii John-ga [tj ageta to]omotteiru (koto).

Mary-toi John-NOM [tj gave that]believe

YUKIKO- 1 -2018/11/17