Two Ways of ĆWide-Scope TakingĻ

J.-R. Hayashishita
University of Southern California

This paper argues that a wide scope reading arises in two distinct ways. It argues in particular that while some cases of the QPi>QPj readings (=WSR1) are based on the LF c-command of the A-position trace of QPj by that of QPi, others (=WSR2) are not. The availability of WSR1 is not affected by the QP type or the syntactic environment in which the relevant sentence occurs, as long as the relevant LF c-command obtains. The availability of WSR2, on the other hand, while not being sensitive to the LF c-command, does get affected by the QP type and the syntactic environment (as well as the pragmatic factors).

In English, the standard generalization (May 1977) is that QP1 VERB QP2 yields both the QP1>QP2 reading (=Q1>Q2) and the QP2>QP1 reading (=Q2>Q1), as illustrated in (1a). In Japanese, on the other hand, the standard generalization (Kuroda 1970, Hoji 1985) is that QP1-ga QP2-CM VERB(where CM = accusative or dative) yields only Q1>Q2. Henceforth Q1 refers to the subject QP and Q2 the object QP.

There are, however, systematic counter-examples to these generalizations. It is pointed out in Liu 1990 and Beghelli & Stowell 1996 that if QP2 cannot be interpreted as a specific group, Q2>Q1 disappears, as illustrated in (1b). Conversely, in Japanese, when QP2 can be interpreted as a specific group such as sanninno syoozyo 'three girls' and subeteno hito 'all people', as opposed to sanninizyoono syoozyo 'more than three girls', kanarinokazuno hito 'a good number of people', 55%izyoono gakusee, 'more than 55% of the students', Q2>Q1 becomes available in addition to Q1>Q2; cf. Kitagawa 1990 and Fox & Sauerland 1995. This is illustrated in (2a); cf. (2b).

Furthermore, the availability of Q2>Q1 hinges on a particular syntactic environment. Hoji (1985) and Yoshimura (1992) report that WCO effects illustrated in (3a) can be remedied when QP-CM 'is fronted' as in (3b). This suggests that the dislocated QP can be in an A-position; cf. Saito 1992. Ueyama (1998), however, points out three environments where a dislocated QP cannot be understood as occupying an A-position. Two of the three environments which are relevant here are i) an embedded clause of a perceptual report construction, illustrated in (4), and ii) when another dislocated QP already manifests such A-properties, illustrated in (5a), contrasting with (5b).

What is of interest is that Q2>Q1 under discussion cannot obtain precisely in the environments where a dislocated QP cannot be understood as occupying an A-position. (6) illustrates that the reading cannot obtain in the perceptual report construction, and (7a) illustrates that the reading cannot obtain when a dislocated QP is to bind soko 'it'. (7a) contrasts with (7b), where a dislocated QP does not bind anything. The availability of Q2>Q1 in English seems to get affected by similar structural considerations. (8) illustrates that if (1a) is placed in a perceptual report construction, Q2>Q1 becomes unavailable.

In sharp contrast to Q2>Q1, Q1>Q2 remains possible regardless of QP types (shown in (9a) for Japanese and (9b) for English) and the syntactic environments (shown in (6) and (7a) for Japanese and in (8) for English). Notice that if QP1 is embedded so that it fails to c-command QP2, the reading disappears as illustrated in (10a) for Japanese and (10b) for English.

What the above discussion shows is that Q1>Q2 can be either an instance of WSR1 or WSR2, while Q2>Q1 must be an instance of WSR2. In retrospect, the Kuroda/Hoji generalization in Japanese can be understood as making reference only to WSR1, while the English generalization makes reference to both WSR1 and WSR2; thus, it seems that the widely-known difference between English and Japanese, noted at the outset, is only apparent. It does not seem to be a mere coincidence that the standard generalization in English has been made mainly with QPs which can be interpreted as a specific group such as everyone and someone; see the earlier generalization in English reported in Chomsky 1957:100. The generalization in Japanese, on the other hand, has been made mainly with QPs which cannot be so interpreted such as sake ka biiru, 'sake or beer' and John-sae 'even John' (Hoji 1985).

In the remainder of the paper, I outline an analysis of WSR2, which incorporates two peculiar properties: i) WSR2 arises when QP is interpreted as a specific group, and ii) WSR2 requires the sentence-initial A-position mentioned above. I will also discuss the relation between WSR2 and Fox & Sauerland's 'illusive scope'. Several other predictions will also be addressed: e.g. Q2>Q1 in (11) is an instance of WSR2 since QP2 does not c-command QP1; thus, the availability of Q2>Q1 is predicted to get affected by the QP type as well as the syntactic environment. The predictions seem to be borne out, as illustrated in (12) and (13).

(1)a  [QP1 two teachers] recommended [QP2 three students].				Q1>Q2, Q2>Q1
   b  [QP1 two teachers] recommended [QP2 more than three students].  		        Q1>Q2, *Q2>Q1         

(2)a  [QP1 yoninno syoonen]-ga [QP2 sanninno syoozyo]-o syootaisite,
                  (sorezoreno syoozyo-ga dareno syootai-o ukerubeki ka mayotteiru).  	Q1>Q2, Q2>Q1
      '[QP1 4 boys] invited [QP2 3 girls],
                  (and each girl is wondering whose invitation she should accept.)'	
   b  [QP1 yoninno syoonen]-ga [QP2 sanninizyoono syoozyo]-o syootaisite,
                  (sorezoreno syoozyo-ga dareno syootai-o ukerubeki ka mayotteiru).	Q1>Q2, *Q2>Q1
      '[QP1 4 boys] invited [QP2 more than 3 girls],
                  (and each girl is wondering whose invitationn she should accept.)'
     
(3)a  *[[ ÷ he ÷ ]NP]-ga QP-CM VERB
      *(kinoono kaigi dewa) sokono bengosi-ga [QP mittuizyoono kaisya]-ni ayamatteita. 
      '(Lit.)(at yesterday's meeting) its attorney was apologizing to [QP more than 3 companies]'
   b  QP-CM [[ ÷ he ÷ ]NP]-ga VERB
	(kinoono kaigi dewa) [QP mittuizyoono kaisya]-ni sokono bengosi-ga ayamatteita
      '(Lit.) (at yesterday's meeting) its attorney was apologizing to [QP more than 3 companies]'
     
(4)   *Karen-DAT [S QP-CM [[ ÷ he ÷ ]NP]-ga VERB] saw  (Cf. (3b).)
      *(party-de) Karen-ni [[QPmittuizyoono kaisya]-ni sokono bengosi-ga ayamatteiru]no-ga mieta
       '(Lit.) (at the party) Karen saw its attorney apologizing to [QP more than 3 companies]'

(5)a  *QP1-CM QP2-CM [[÷ he2 ÷ he1 ÷ ]NP]-ga VERB
      *[QP1USC to UW]1-o[QP2 sanninno gakusee]2-ni [[soitu2-o sitteiru]soko1no sensee]-ga suisensaseta.
	'(Lit.) [its1 professor who knows him2] made [QP2 3 students]2 recommended [QP1USC and UW]1'
   b  QP1-CM QP2-CM [[÷ John ÷ he1 ÷ ]NP]-ga VERB
      [QP1USC to UW]1-o [QP2 sanninno gakusee]-ni [[John-o sitteiru soko1no sensee]-ga suisensaseta.
	'(Lit.) [its1 professor who knows John] made [QP2 3 students] recommended [QP1 USC and UW]1'

(6)   John and Bill (respectively)-DAT [S QP1-NOM QP2-CM VERB] saw  (Cf. (2a) and (4).)
      John to Bill sorezoreni [QP1yoninno syoonen]-ga [QP2sanninno syoozyo]-o syootaisiteiru no-ga mieta
      'John and Bill respectively saw [QP1 4 boys] inviting [QP2 3 girls]. 		Q1>Q2, *Q2>Q1

(7)a  QP-CM [[÷ it÷]QP1]-ga QP2-CM VERB  (Cf. (5a).)				Q1>Q2, *Q2>Q1
      [QP3USC to UW]-o[QP1soko-o sotugyoosita [hutarino sensee]]-ga [QP2sanninno gakusee]-ni suisensita.
	'(Lit.) [QP1 2 teachers who graduated from it] recommended [QP3USC and UW] to [QP2 3 students]' 
   b  QP-CM [[÷ UT÷]QP1]-ga QP2-CM VERB  (Cf. (5b).)	     			Q1>Q2, Q2>Q1
      [QP3USC to UW ]-o[QP1UT-o sotugyoosita [hutarino sensee]]-ga [QP2sanninno gakusee]-ni suisensita.
	'[QP1 2 teachers who graduated  from UT] recommended [QP3 USC and UW] to [QP2 3 students].'

(8)   John and Mary (respectively) saw [QP1two teachers] recommending [QP2three students]. 
									            (Cf. (1a).) Q1>Q2,*Q2>Q1

(9)a  [QP1yoninizyoono syoonen]-ga [QP2sanninizyoono syoozyo]-o syootaisite,
             (sorezoreno syoozyo-ga dareno syootai-o ukerubeki ka mayotteiru).     (Cf. (2b).) Q1>Q2, *Q2>Q1
      [QP1 more than 4 boys] invited [QP2 more than 3 girls],
             (and each girl is wondering whose invitation she should accept.)'
   b  [QP1 more than 2 teachers] recommended [QP2 more than 3 students].           (Cf. (1b).) Q1>Q2, *Q2>Q1

(10)a [ ÷ QP1 ÷ ]-ga QP2-CM VERB  (The translation of (10a) is given in (10b).) 	     *Q1>Q2, *Q2>Q1 
      [John-ga [QP1 hutariizyoono onna]-to kisusita koto]-ga [QP2 sanninizyoono hito]-o situboosaseta.
    b [That John kissed [QP1 more than two women]] disappointed [QP2 more than three people].

(11)  At the party Bill greeted [QP1 someone] from [QP2 three cities].			       Q1>Q2 Q2>Q1

(12)  At the party Bill greeted [QP1 someone] from [QP2 more than three cities].	       Q1>Q2 *Q2>Q1

(13)  John and Mary (respectively) saw Bill greeting [QP1 someone] from [QP2 three cites].     Q1>Q2 *Q2>Q1

References

Fox, Danny & Uli Sauerland (1995) "Illusive Scope of Universal Quantifiers", NELS 26

Kitagawa, Yoshihisa (1990) "Anti-Scrambling," ms. University of Rochester.

Ueyama, Ayumi (1998) Two Types of Dependency, Doctoral dissertation, USC

Yoshimura, Noriko (1992) Scrambling and Anaphora in Japanese, Doctoral dissertation, USC


< Back to List of Abstracts | Back to PLC23 Home Page >

About the PLC23 Committee
Previously held Penn Linguistics Colloquium: PLC22 (1998), PLC21 (1997)

Penn Department of Linguistics
University of Pennsylvania