To appear in Japanese/Korean Linguistics 18, CSLI Publications, Stanford, California. Prefinal version.

Degree Quantification and the Size of Noun Modifiers JUNKO SHIMOYAMA* McGill University

1. Direct vs. Indirect Modification Direct attributive uses of adjectives as in (1) have often been considered a characteristic property of adjectives. Recent studies suggest, however, that some languages lack direct modification by adjectives entirely and that they only allow for indirect modification using relative clauses as in (2). (1) [AP expensive] flowers (2) a. flowers [CP which1 [TP are e1 expensive]] b. flowers [CP which1 [TP John bought e1]]

*I would like to thank the audience at the 18th Japanese/Korean Linguistics Conference at CUNY for useful comments. Earlier versions of the paper also benefited from comments from the audiences at University of Texas at Austin, University of Massachusetts at Amherst, McGill University, and Concordia University. The work reported here was supported in part by a startup grant from McGill University and by a grant from FQRSC (2008-NP-121129), which I gratefully acknowledge. All remaining errors belong to me.

For example, Baker (2003) points out that Slave is a language of this type. Example (3a) contains a predicative use of the adjective sho ‘proud/happy’, while (3b), attributed to Keren Rice (p.c.), shows that this adjective cannot be used attributively. Though there is no example, it is mentioned in the text (Baker 2003: 194) that indirect modification with a relative clause as in a woman that is proud must be used to express this idea. (3) a. b.

Yenene (be-ghǫ) sho hįlį woman 3-of proud/happy 3-is ‘The woman is proud/happy (of him/her).’ *yenene [sho] woman proud/happy ‘a proud/happy woman’

(Rice 1989)

Although less obvious on the surface, Korean has also been claimed to lack direct modification by adjectives. According to M.-J. Kim (2002), what may seem like a direct adjectival modifier in (4b) is arguably an indirect relative clause modifier. (4) a. b. c.

ce yeca-ka yeppu-ta that woman-NOM pretty-DCL ‘That woman is pretty. [e yeppu]-n yeca pretty-REL woman ‘The/a woman who is pretty’ [e ttena]-n namca left-REL man ‘the/a man who left’

The idea that some apparent attributive uses of adjectives have predicative sources in relative clauses has also been proposed for English and other languages (e.g. Smith 1961, 1964, Ross 1967, Jacobs and Rosenbaum 1968, Sproat and Shih 1991, Kayne 1994, Larson 1999, Alexiadou 2001). For example, the sentences that contain prenominal and postnominal adjectival modifiers in (5) from Larson (1999) could be derived from the underlying sentence in (6), by way of relative clause reduction (WHIZ deletion, Ross 1967) and optional movement of the adjective. (5) a. The visible stars include Capella. b. The stars visible include Capella. (6) The stars [which are visible] include Capella. 2

2. Japanese Adjectival Modifiers and a Relative Clause Parse At a first glance, Japanese seems to fit into the group of languages that lack direct adjectival modification, as suggested by Baker (2003). This is in fact a widely accepted view on Japanese adjectival modification (e.g. Kuno 1973, Shibatani 1978, Whitman 1981, Dixon 1982, Miyagawa 1984, Makino and Tsutsui 1986, Urushibara 1993, Nishiyama 1999, 2005). Unlike their English counterparts, Japanese adjectival modifiers as in (7) are consistent with a relative clause parse in (8).1 This is due to three factors. First, noun modifiers are always prenominal in Japanese, which is not the case in English as we can see in (1) and (2). Second, Japanese lacks overt relative pronouns. Finally, though the status of the morpheme -i is unclear, it has standardly been assumed to be a present tense morpheme or a presenttensed form of copula (see the references above as well as Yamakido 2005). (7) (8) (9)

[utukusi-i] hana beautiful-I flower ‘beautiful flowers’ [CP Op1 [TP e1 utukusi-i]] hana beautiful-I flower ‘flowers that are beautiful’ [CP Op1 [TP John-ga e1 kat-ta]] hana John-NOM buy-PAST flower ‘flowers that John bought’

There is some suggestive evidence that the relative clause parse in fact exists for apparent attributive uses of adjectives in Japanese. Sproat and Shih (1991) observe that in Mandarin Chinese, while adjectival modifiers without de in (10) obey the adjective ordering restriction observed in many languages in (12), adjectival modifiers with de do not, as seen in (11) (see also Paul 2005). (10)

(11)

a.

xiao lü huaping small green vase b. ? lü xiao huaping green small vase a. xiao-de lü-de small-DE green-DE b. lü-de xiao-de green-DE small-DE

SIZE > COLOR COLOR > SIZE

huaping vase huaping vase

1The question of whether Japanese relative clauses are CPs or TPs does not directly affect the main points in this paper (see Murasugi 1991).

3

(12)

‘Basic’ order of adjectives: QUALITY > SIZE > SHAPE > COLOR > PROVENANCE (e.g. Whorf 1945, Dixon 1982, Sproat and Shih 1991 and the references cited there.)

Sproat and Shih (1991) propose that adjectival modifiers with de should receive a relative clause analysis. This fits with the observation that no ordering restriction is observed in English relative clauses. The following examples are from Larson and Takahashi (2007). (13) (14)

a. the large stone building b. * the stone large building a. the building [that was large][that was made of stone] b. the building [that was made of stone][that was large]

The ordering facts in Japanese in (15) below makes it all the more plausible that Japanese adjectival modifiers are part of relative clauses. (15)

a. b.

ooki-i maru-i onigiri large-I round-I rice-ball ‘large round rice-ball’ maru-i ooki-i onigiri round-I large-I rice-ball ‘round large rice-ball’

Since we have not seen any evidence against a relative clause parse, I will assume for this paper that it is available. That is, we have seen nothing that prevents the parse in (16a) for, for instance, the example in (7). (16)

a. [CP Op1 [TP e1 is [AP A]]] NP b. [AP A] NP

The question now is whether direct modification by adjectives in (16b) exists on top of the indirect modification by relative clauses. The question is an important one because, if no additional assumptions are made, one would expect direct adjectival modification to be available. If it in fact turns out that Japanese lacks direct adjectival modification, then one would want to know why. Baker (2003), for example, proposes that in some languages, including Japanese, APs cannot stand alone and need to be supported by PredP (predication phrase). I am not aware of any deeper explanation where a lack of direct modification follows naturally. 4

In what follows, I will present evidence that direct adjectival modification without the support of finite TP and CP layers exist in Japanese. Though syntactic evidence for the presence of direct modifiers is hard to come by, I show that the scope of degree quantifiers can be used as a probe into the structure of direct vs. indirect modifiers. The results indicate that adjectival modifiers and clear cases of relative clauses do not show the same range of interpretations, suggesting that after all, adjectival modifiers in Japanese cannot all receive a relative clause parse. This provides further support for works by Kusumoto (1999) and Yamakido (2005), which look at, among other things, the temporal interpretation of adjectival modifiers and compare that with clear cases of relative clauses, and show that they behave differently.

3. Scope Island Effect in Noun Modifiers 3.1. Scope Islands for Degree Quantifiers In examining the size of noun modifiers in Japanese, we look at the availability of a so-called comparative reading of superlative noun phrases as in (17b) (Ross 1964, Jackendoff 1972, Szabolcsi 1986, Heim 1985, 1999, among others). (17)

John climbed the highest mountain. (Szabolcsi 1986) a. Absolute reading John climbed a mountain that is higher than any other mountain that is contextually relevant. b. Comparative reading John climbed a higher mountain than anyone else did.

While the absolute reading compares the height of mountains, the comparative reading compares the relevant mountain climbers’ achievements. In the latter reading, the sentence asserts that John ranks highest in terms of how high a mountain he climbed. Comparative readings are said to be sensitive to scope islands (Szabolcsi 1986). Let us consider the examples in (18), which are based on Szabolcsi (1986).2 In (18a), the superlative morpheme -est occurs inside a nonfinite complement clause, and in (18b), it occurs inside a finite complement clause. Sentence (18a) has two comparative readings, low and high, paraphrased in (19). In the low comparative reading in (19a), John’s desire is to outscore everyone. In the high comparative reading in (19b), John is the most ambitious one in terms of how high a score one desires to get. Sen2See also Farkas and Kiss (2000).

5

tence (18b), on the other hand, only has the low comparative reading, paraphrased in (20a). It lacks the high comparative reading in (20b). (18) a. John wants [PRO to get the highest score]. (non-finite complement) b. John said [that he got the highest score]. (finite complement) (19) a. Low comparative reading for (18a) What John wants is to get a higher score than anyone else. b. High comparative reading for (18a) The score that John wants to get is higher than the score anyone else wants to get. (20) a.

Low comparative reading for (18b) John said that he got a higher score than anyone else got. b. * High comparative reading for (18b) The score that John said he got is higher than the score anyone else said John (or he/she) got.

The contrast in (18a) and (18b) can be derived if we assume that comparative readings come about via movement of the superlative morpheme est (see Heim 1999 for details).3 According to a movement analysis, sentence (21a) in its comparative reading has the LF structure in (21b). Combined with the lexical semantics for -est in (21c), it gives you the truth conditions in (21d), which says that there is a degree d such that John climbed a d high mountain and no one else in the contextually determined set climbed a d high mountain. (21) a. b. c. d.

John climbed the highest mountain. John -estC 1 climbed [a [t1,d high] mountain] [[-est]] = λC.[λf.[λxe: x∈C. ∃d[f(d)(x) = T & ∀y[y∈C & y≠x → f(d)(y) = F] ] (C = the set of relevant people) ∃d[John climbed a d high mountain & ∀y[y ∈ C & y ≠ x → y did not climb a d high mountain]]

In a movement analysis, we can understand the lack of the high comparative reading in (18b) as a scope island effect. Finite clauses, but not non-finite clauses, create islands for the movement of the superlative morpheme -est.

3See also Sharvit and Stateva (2002).

6

3.2. The Lack of Scope Island Effect in Japanese Adjectival Modifiers Going back to Japanese noun modifiers, recall that we are interested in the question of whether direct adjectival modification exists in addition to indirect relative clause modification. The first test that comes to mind, faced with this kind of question, may be one that checks whether wh-extraction (in such constructions as wh-interrogatives or relative clause formation) is possible out of such noun modifiers, as relative clauses are known to create islands for syntactic extraction. There are a few reasons why we cannot rely on this type of standard test in the context of Japanese. First, it is not clear whether wh-movement exists in wh-interrogatives and relative clauses. Second, it is not clear whether relative clauses form islands for movement. Furthermore, even without these Japanese-specific considerations, it is not easy to set up examples in which wh-movement takes place out of an adjectival modifier, leaving an individual trace. This is why looking at a degree quantifier like -est, which specializes in degree predicates such as adjectives, would help here. I assume that the Japanese superlative morphemes mottomo and itiban in (22) are the counterparts of -est or most in English, and that they share the meaning in (21c). Unlike motto ‘more’, which is often described as an optional comparative morpheme, mottomo/itiban are obligatory. (22) Hanako-wa kumi-de/ gakusei-no naka-de *(mottomo/itiban) Hanako-TOP class-in/ student-GEN among most kashiko-i. smart-I ‘Hanako is smartest in class/among students.’ Consider now the sentence in (23) and the scenario in (25). Suppose someone utters the sentence while pointing to triangle 1. The sentence in its absolute reading is judged false in this scenario. The sentence, however, has another reading in which it is judged true in the same scenario. That is the comparative reading in (24). (23) Kono sankakkei-ga [[ mottomo/itiban ooki-i] en]-ni sesssteiru. this triangle-NOM most large-I circle-DAT touch ‘This triangle touches the largest circle.’ (24) ∃d[this triangle touches a d large circle & ∀y[y ∈ C & y ≠ x → y does not touch a d large circle]] (C = the contextually determined set of triangles)

7

(25) 2

1

3

In its comparative reading, the English sentence in (26a) is judged to be true in the same scenario. On the other hand, sentence (26b) lacks this reading and only has the absolute reading, thus the sentence is judged to be false in this scenario. (26) a. b.

Triangle 1 touches the [largest] circle. Triangle 1 touches the circle [that is largest].

The contrast is expected, given what we saw in (18) above. The superlative morpheme -est cannot be extracted out of finite clauses. Thus, the Japanese noun modifier ooki-i ‘large-I’ in (23) behaves more like large in English than the relative clause that is large.4 This would receive a natural explanation if we adopt two assumptions. First, the comparative reading in the Japanese sentence in (23) is derived by movement of the superlative morpheme mottomo or itiban to the matrix VP. Aihara (2007) in fact provides arguments for this view. Second, direct adjectival modification is available in Japanese. (27) a. [This triangle] [mottomo1 [a [t1,d large] circle] touches] b. [This triangle] [mottomo1 [a [that [TP is t1,d large]] circle] touches] At this point, it is important to show that superlative noun phrases in Japanese do exhibit scope island effects elsewhere. When the superlative morpheme needs to move out of a clearly finite relative clause as in (28b) and (29b), the comparative reading becomes unavailable. 4In both Korean and Japanese, the comparative reading is not available when a topic marker is used, unless it is interpreted contrastively. (i) John{-i/-un} gajang nopun sanul ollassda. (Korean, Lyu 2004) John-NOM/-TOP most high mountain-ACC climbed ‘John climbed the highest mountain.’ Lyu’s (2004) tentative conclusion is that abstract focus marking is necessary for the center of comparison, and topic marked phrases are incompatible with abstract focus marking. In Hungarian, focal stress is necessary for (overt) center of comparison (see Szabolcsi 1986 and Farkas and Kiss 2000).

8

(28) a. Taro-ga [[ mottomo/itiban omosiro-i] hito]-o syuzaisita. Taro-NOM most funny-I person-ACC interviewed ok ‘Taro interviewed the funniest person.’ Comparative b. Taro-ga [[ mottomo/itiban omosirok-at-ta] hito]-o Taro-NOM most funny-COP-PAST person-ACC syuzaisita. *Comparative interviewed ‘Taro interviewed the person who was/had been funniest.’ (29) a. Ken-ga [[mottomo/itiban sinsen-na] sakana]-o hurumatta. Ken-NOM most fresh-NA fish-ACC treated ok ‘Ken treated (the guests) to the freshest fish.’ Comparative b. Ken-ga [[mottomo/itiban sinsen-dat-ta] sakana]-o hurumatta. Ken-NOM most fresh-COP-PAST fish-ACC treated ‘Ken treated (the guests) to the fish that had been freshest.’ *Comparative Furthermore, the Japanese counterparts to the English sentences in (18) behave in the same way. That is, mottomo/itiban can move out of non-finite complements, but not out of finite complements.

4. Conclusions and Remaining Questions We have seen that Japanese adjectival modifiers do not show the scope island effect for a superlative degree quantifier that is expected under the standard view that they are present-tensed relative clauses. I take this as evidence that direct modification exists in Japanese. This conclusion provides further support for works by Kusumoto (1999) and Yamakido (2005), which look at, among other things, the temporal interpretation of adjectival modifiers and compare that with clear cases of relative clauses, and show that they behave differently. If adjectives in Japanese do form direct modifiers, no stipulation is necessary as to how to make sense of the apparent lack of direct modification. There is also a need now to reexamine claims made for other languages such as Korean and Mandarin Chinese. Furthermore, if the story presented here is on the right track, the morpheme -i (and -na) must be dissociated from present tense. When adjectives that end with these morphemes occur in full clauses, then, perhaps there is a null present tense. Yamakido (2005, 2007) presents suggestive evidence for dissociation of -i and tense from the Wakayama dialect of Japanese, in which -i and the past tense -ta cooccur (see also Larson and Yamakido 2008). The study reported here opens up many interesting questions to be explored. What I have excluded above as a source of comparative reading is indirect modification with tense. The possibility of indirect modification that does not come with tense needs to be examined. The diagnostics used 9

in this paper relying on the scope of degree quantifiers may also help detect structure in reduced relatives such as (30b) (see, for example, Bhatt 2006 and Belikova 2008).5 (30) a. John ate the doughnut [that was touched by the fewest people]. b. John ate the doughnut [touched by the fewest people]. Looking at so-called adjectival relatives in Japanese is an obvious next step to pursue (Kusumoto 2002 and Ogihara 2004). Example (31) seems to allow a comparative reading, despite the occurrence of what appears to be the past-tense morpheme -ta in the modifier. (31) Haruko-ga [[ mottomo/itiban magat-ta] miti]-o hasit-ta. Haruko-NOM most wind-TA road-ACC ran-PAST ‘Haruko ran the most winding road.’ Interestingly, both Kusumoto (2002) and Ogihara (2004) propose that adjectival relatives have a more reduced structure than full relative clauses. Furthermore, Miyagawa (2008) observes that in sentence (32), when the genitive case is used, the non-eventive interpretation in (32a) is strongly preferred over the eventive interpretation in (32b), associated with a full relative clause structure. He suggests that the genitive case shows up when the modifier has a reduced structure, AspP.6 (32) [Kizu-ga/no tui-ta] kabe scratch-NOM/GEN stick-TA] wall a. ‘the scratched wall’ b. ‘the wall that was scratched’ This may suggest that the availability of comparative readings would correlate with the occurrence of the genitive case in adjectival relatives. It is not immediately obvious that this correlation holds. For Kusumoto (2002), on the other hand, modification structure is reduced regardless of which case marking, nominative or genitive, is used in examples like (32). If we could find cases where the use of the nominative case uniquely corresponds to an eventive interpretation, and the use of genitive case to a non-eventive interpretation, it would be interesting to see whether comparative readings would be limited to the latter environment.7 Further work is needed. 5I thank Heidi Harley (p.c.) for the examples in (30). 6I thank an anonymous J/K abstract reviewer for bringing Miyagawa (2008) to my attention. 7Thanks to Satoshi Tomioka for this point.

10

References Aihara, M. 2007. Japanese Superlative Constructions: Evidence for ‘Est’-movement. Proceedings of the 37th Annual Meeting of the North East Linguistic Society (NELS 37). GLSA, University of Massachusetts, Amherst. Alexiadou, A. 2001. Adjective Syntax and Noun Raising: Word Order Asymmetries in the DP as the Result of Adjective Distribution. Studia Linguistica 55:217-248. Baker, M. C. 2003. Lexical Categories: Verbs, Nouns, and Adjectives. Cambridge: Cambridge University Press. Belikova, A. 2008. Syntactically Challenged Rather Than Reduced: Participial Relatives Revisited. Proceedings of the 2008 Annual Conference of the Canadian Linguistic Association. Bhatt, R. 2006. Covert Modality in Non-finite Contexts. Berlin: Walter de Gruyter. Dixon, R. M. W. 1982. Where Have All the Adjectives Gone? Berlin: Mouton. Farkas, D. F. and K. E. Kiss. 2000. On the Comparative and Absolute Reading of Superlatives. Natural Language and Linguistic Theory 18:417-455. Heim, I. 1985. Notes on Comparatives and Related Matters. Ms., University of Texas at Austin. Heim, I. 1999. Notes on Superlatives. Ms., MIT. Heim, I. 2000. Degree Operators and Scope. Semantics and Linguistic Theory X, pp. 40-64. Kim, M.-J. 2002. The Absence of Adjectives and Noun Modification in Korean. Proceedings of the 2002 International Conference on Korean Linguistics, Seoul: Hankwuk Mwunhwasa. Kuno, S. 1973. The Structure of the Japanese Language. Cambridge, Mass.: MIT Press. Kusumoto, K. 1999. Tense in Embedded Contexts. Doctoral dissertation, University of Massachusetts, Amherst. Kusumoto, K. 2002. The Semantics of the Non-past -ta in Japanese. Proceedings of the Third Formal Approaches to Japanese Linguistics, 163-180. MIT Working Papers in Linguistics 41. Larson, R. 1999. Semantics of Adjectival Modification. Lecture notes, LOT Winter School, Amsterdam. Larson, R. and N. Takahashi. 2007. Order and Interpretation in Prenominal Relative Clauses. Proceedings of the Second Workshop on Altaic Formal Linguistics, ed. M. Kelepir and B. Öztürk, MIT Working Papers in Linguistics 54. Larson, R. and H. Yamakido. 2008. Ezafe and the Deep Position of Nominal Modifiers. Adjectives and Adverbs: Syntax, Semantics, and Discourse, ed. L. McNally and C. Kennedy, 43-70. Oxford: Oxford University Press. Lyu, H.-Y. 2004. Focus and Superlative Center in Korean. Handout for LIN 393: Seminar on Noun Phrase Modification, University of Texas at Austin.

11

Makino, S. and M. Tsutsui. 1986. A Dictionary of Japanese Grammar. Tokyo: Japan Times. Miyagawa, S. 1984. Lexical Categories in Japanese. Lingua 73: 29-51. Miyagawa, S. 2008. Genitive Subjects in Altaic. Proceedings of the Workshop on Altaic Formal Linguistics 4, MIT Working Papers in Linguistics 56. Murasugi, K. 1991. Noun Phrases in Japanese and English: A Study in Syntax, Learnability and Acquisition. Doctoral dissertation, University of Connecticut. Nishiyama, K. 1999. Adjectives and the Copulas in Japanese. Journal of East Asian Linguistics 8: 183-222. Nishiyama, K. 2005. Morphological Boundaries of Japanese Adjectives: Reply to Namai. Linguistic Inquiry 36: 134-142. Ogihara, T. 2004. Adjectival Relatives. Linguistics and Philosophy 27: 557-608. Paul, W. 2005. Adjectival Modification in Mandarin Chinese and Related Issues. Linguistics 43: 757-793. Ross, J. R. 1964. A Partial Grammar of English Superlatives, MA thesis, University of Pennsylvania. Ross, J. R. 1967. Constraints on Variables in Syntax. Doctoral dissertation, MIT. Sharvit, Y. and P. Stateva. 2002. Superlative Expressions, Context, and Focus. Linguistics and Philosophy 23: 453-504. Shibatani, M. 1978. Nihongo no bunseki (Analysis of Japanese). Tokyo: Taishukan. Smith, C. S. 1961. A Class of Complex Modifiers in English. Language 37: 342365. Smith, C. S. 1964. Determiners and Relative Clauses in a Generative Grammar of English. Language 40: 37-52. Reprinted in Modern Studies in English, ed. D. A. Reibel and S. A. Schane (1969) 247-263. Englewood Cliffs, N. J.: Prentice-Hall. Sproat, R. and C. Shih. 1991. The Cross-Linguistic Distribution of Adjective Ordering Restrictions. Interdisciplinary Approaches to Language: Essays in Honor of S.-Y. Kuroda, ed. C. Georgopoulos and R. Ishihara, 565-593. Dordrecht: Kluwer. Szabolcsi, A. 1986. Comparative Superlatives. MIT Working Papers in Linguistics 8, ed. Naoki Fukui et al., 245-265. Urushibara, S. 1993. Syntactic Categories and Extended Projections in Japanese. Doctoral dissertation, Brandeis University. Whitman, J. 1981. The Internal Structure of NP in Verb Final Languages. Proceedings of the Chicago Linguistic Society 17, 411-418. Yamakido, H. 2000. Japanese Attributive Adjectives are not (all) Relative Clauses. Proceedings of WCCFL 19, 588-602. Yamakido, H. 2005. The Nature of Adjectival Inflection in Japanese. Doctoral dissertation, Stony Brook University. Yamakido, H. 2007. The Nature of Adjectival Inflection in Japanese. Proceedings of the Second Workshop on Altaic Formal Linguistics, ed. M. Kelepir and B. Öztürk, MITWPL 54. 12