1 Ordering of Attributive Adjectives

Sproat and Shih (1991) characterize Japanese as a language that uses indirect modification for adnominal adjectives, on the basis of the observation that adjectival ordering is free, as in (1).

(1)

  • a.

    chiisana shikakui ie

    small square house

  • b.

    shikakui chiisana ie

    square small house

    ‘small square house’

    (Sproat and Shih 1991:582)

In contrast, direct modification is exemplified by prenominal adjectives in English, which exhibit rigid ordering, as in (2).

(2)

Building on a long history of research on adjectival ordering summarized in Sproat and Shih 1991 and incorporating further data, Scott (2002) proposes the following hierarchy of attributive adjectives, which characterizes direct modification structure:

(3)

  • Hierarchy of attributive adjectives

    subjective comment > ?evidential > size > length > height > speed > ?depth > width > weight > temperature > ?wetness > age > shape > color > nationality/origin > material

    (Scott 2002:114)

The contrast in (2) is attributed to the relative height of size and shape adjectives in the hierarchy. The hierarchy is encoded as the layer of functional heads, shown schematically in (4), that host the relevant class of adjectives in their specifier, as proposed by Cinque (1994).

(4)

graphic

Under this proposal, there is a functional head corresponding to each adjectival class. See Aljović 2010 for evidence concerning the morphological realization of these functional heads.

Direct modification is to be distinguished from indirect modification in other respects as well. Sproat and Shih (1991) observe that direct modifiers are closer to the head noun than indirect modifiers: they point to the contrast in (5) in Mandarin, where indirect modifiers are marked by de while direct modifiers occur without it.

(5)

  • a.

    hei-de xiao shu

    black-DE small book

    ‘small black book’

  • b.

    *xiao hei-de shu

    small black-DE book

    (Sproat and Shih 1991:571)

The indirect modifier status of de-marked phrases can be seen from the free ordering in (6).

(6)

  • a.

    xiao-de lü-de huaping

    small-DE green-DE vase

  • b.

    lü-de xiao-de huaping

    green-DE small-DE vase

    ‘small green vase’

    (Sproat and Shih 1991:565)

Bare modifiers, on the other hand, exhibit rigid ordering as in (7).

(7)

  • a.

    xiao lü huaping

    small green vase

    ‘small green vase’

  • b.

    *lü xiao huaping

    green small vase

    (Sproat and Shih 1991:566)

Thus, direct modifiers are doubly hierarchical, within their own category as well as vis-à-vis indirect modification.

Furthermore, relative clauses cannot be used as direct modifiers, as shown by the fact that a relative clause cannot appear in prenominal position in English.

In this squib, I will show that direct modification structure is found in Japanese as well, giving rise to rigid ordering of some classes of modifiers, and furthermore that nouns as well as adjectives can participate in direct modification.

2 Classes of Attributive Modifiers in Japanese

What concerns us in the discussion of Japanese is the portion of (3) having to do with nationality/origin and material. The relative height of this portion of the hierarchy vis-à-vis other types of adjectives in English is illustrated in (8) and (9).

(8)

(9)

When we turn to Japanese, the first thing to note is that the class of adjectives that express nationality/origin and material is absent. Instead, nouns are used as modifiers, as shown in (10).

(10)

  • a.

    chuugoku-no kabin

    China-GEN vase

    ‘Chinese vase’

  • b.

    ki-no hashi

    wood-GEN bridge

    ‘wooden bridge’

The crucial evidence for the nominal nature is the presence of no attached to the modifier noun (Kitagawa and Ross 1982, Nishiyama 1999).1 Genuine adjectives do not use no, as illustrated in (11). See also (1) above.

(11)

  • nagai hashi

  • long bridge

  • ‘long bridge’

Note also that adjectives of nationality/origin and material are denominal in English.

Let me also point out that it is not the case that Japanese lacks denominal adjectives altogether. There are suffixes like -ppoi that create adjectives out of nouns, as illustrated in (12).

(12)

  • a.

    hokori-ppoi

    dust-SUFFIX

    ‘dusty’

  • b.

    mizu-ppoi

    water-SUFFIX

    ‘watery’

The nonnominal status of such forms is shown by the absence of no in modification structures, as in (13).

(13)

  • hokori-ppoi michi

  • dust-SUFFIX road

  • ‘dusty road’

Thus, it is an important characteristic of Japanese that denominal adjectives of nationality/origin and material do not exist.

Now we are ready to examine the ordering restriction on attributive modifiers in Japanese. Note the following contrast:

(14)

  • a.

    chiisana ki-no hashi

    small wood-GEN bridge

    ‘small wooden bridge’

  • b.

    ??ki-no chiisana hashi

    wood-GEN small bridge

(15)

  • a.

    chiisana chuugoku-no kabin

    small China-GEN vase

    ‘small Chinese vase’

  • b.

    ??chuugoku-no chiisana kabin

    China-GEN small vase

The preferred position of size adjectives is higher than the positions of modifiers expressing nationality/origin or material. In evaluating the contrast, it is important to avoid inserting a pause after the modifiers. (14a) and (15a) are acceptable without a pause. (14b) and (15b), judged ?? without pauses, become perfectly acceptable if the speaker pauses after the no-marked modifier. In that case, a wooden bridge is contrasted with, say, a steel bridge in (14b). Similar remarks apply to (15b). As noted by Scott (2002), focused adjectives can be preposed to a designated higher position within DP in English through phrasal focus movement (pace Cinque 2010:59). It is not surprising to find essentially the same process in Japanese. Thus, the version of (14b) and (15b) with a pause should be derived from (14a) and (15a), respectively, by raising the modifying noun phrase as in (16).

(16)

  • a.

    ki-noi, chiisana ti hashi

  • b.

    chuugoku-noi, chiisana ti kabin

We are led to conclude that once these discourse-information factors are suppressed, the ordering restriction on modifiers of nouns is revealed in Japanese as well.2

Though not included in the hierarchy discussed in the literature, adjectives of weather such as rainy and stormy do not exist in Japanese, either. It may not be a coincidence that they are again denominal in English. In any case, nouns expressing weather must be closer to the modified noun than adjectives in Japanese, as shown in (17).3

(17)

  • a.

    shizukana yuki-no hi

    quiet snow-GEN day

    ‘quiet snowy day’

  • b.

    *yuki-no shizukana hi

    snow-GEN quiet day

At this point, one might entertain the possibility that as modifiers of nouns, adjectives occur outside the domain that hosts modifier nouns. The source of the word order restriction is not the categorial distinction between nouns and adjectives, however. There is an ordering between nationality/origin and material, too, in conformity with the hierarchy in (3), as illustrated in (18) and (19).

(18)

  • a.

    chiri-no kin-no kubikazari

    Chile-GEN gold-GEN necklace

    ‘Chilean gold necklace’

  • b.

    *kin-no chiri-no kubikazari

    gold-GEN Chile-GEN necklace

(19)

  • a.

    hokuoo-no ki-no isu

    North.Europe-GEN wood-GEN chair

    ‘North European wooden chair’

  • b.

    *ki-no hokuoo-no isu

    wood-GEN North.Europe-GEN chair

Since these modifiers are both nouns, the categorial distinction has nothing to say about the word order restriction illustrated in (18) and (19). We thus need to recognize the effects of the modifier hierarchy in Japanese.4

This recognition brings us back to cases like (1), repeated here, where adjectives can be freely ordered.

(1)

  • a.

    chiisana shikakui ie

    small square house

  • b.

    shikakui chiisana ie

    square small house

    ‘small square house’

    (Sproat and Shih 1991:582)

My suggestion is that as originally claimed by Sproat and Shih (1991), these adjectives can belong to indirect modification structure, which does not impose a rigid order. The hierarchy in (3) is a feature of direct modification. Since indirect modification is found outside the domain of direct modification, in (14) and (15) these adjectives precede nationality/origin or material modifiers, which are confined to the direct modification region. This analysis should apply to (17) as well.

It is an open empirical question whether size and shape adjectives in (1) and other examples can be used for direct modification in Japanese. As far as word order phenomena are concerned, there is nothing wrong with recruiting them for direct modification. The possibility of indirect modification simply obliterates the crucial word order effect. One relevant consideration is whether Japanese adjectives can be used as noun modifiers without being placed in relative clauses (see Baker 2003b and Yamakido 2005 for opposing views), since relative clauses cannot enter into direct modification structure. At the same time, confinement to indirect modification does not mean that the relativization strategy in finite clauses needs to be used for adjectives to modify nouns, since indirect modification is not limited to full relative clauses (Cinque 2010).5 Future research must sort out these questions.

3 Nouns as Adnominal Modifiers?

We have seen that the lowest portion of the direct modification structure involves nouns rather than adjectives in Japanese. This fact poses a serious challenge to Baker’s (2003a,b) theory of lexical categories, which prohibits nouns from functioning as attributive modifiers. Simplifying somewhat, Baker’s theory says that a noun has a referential index that must be linked to a θ-role. If a noun modifies another noun in attributive structure, the referential index of one of these nouns cannot be linked to a θ-role, resulting in ungrammaticality. The Japanese data call this theory into question.

English presents potentially problematic cases, too. The word order effect is found with material nouns, as illustrated by the following contrast:

(20)

One might say that expressions like stone building are compounds, which would be sufficient to rule out (20b). However, Giegerich (2004) has shown that the compound analysis is untenable. He gives several arguments, but the most telling is the observation that one-anaphora is possible, as in (21).

(21) a wooden bridge and a steel one

Thus, the structure in question is phrasal in nature.6 Giegerich observes further that English has no derived adjectives for the vast majority of material nouns. Adjectives like wooden form a small class. Thus, cases like (20) and (21) are also problematic for Baker’s theory.

One can put aside the English data by saying that these material nouns are turned into adjectives by a null affix. This is indeed the possibility suggested but rejected by Liberman and Sproat (1992), on the grounds that it is not obvious how to restrict the distribution of this affix. Liberman and Sproat’s objection loses its force, though, once we recall that various classes of adjectives must be recognized in any case and that the distribution of the null affix can be restricted accordingly.

Solutions of this sort are not available for the Japanese problem. There is a sharp morphological distinction between nouns and adjectives when they are used as attributive modifiers, as noted above. Remember that modifier nouns need to be followed by a signature marker no, whereas adjectives are not. If the Japanese data are taken at face value, the theory of Universal Grammar must allow nouns to function as attributive modifiers, contrary to Baker’s theory, which in turn leaves the treatment of material nouns in English attributive structure completely open again.

4 Conclusion

Though recent research on Japanese has shown that relative clauses expressing individual-level properties appear closer to the head noun than those expressing stage-level properties (Larson and Takahashi 2007) and that the Japanese counterpart of complete idiot is a unit that cannot be broken up (Ayano 2010),7 the relevance of the hierarchy in (3) has not been explored. I have indicated in this squib that direct modification structure corresponding to the lowest portion of the hierarchy in (3) involving nationality/origin and material is found in Japanese as well. In addition, I have shown that the classification of adjectives on which (3) is based is effective in approaching parametric variation concerning the syntax of adnominal modifiers.

One significant parametric characteristic of Japanese adnominal modifiers is that denominal adjectives of nationality/origin and material are absent. Nouns can directly participate in modification structure. Scott (2002) claims that functional projections hosting adnominal modifiers in a structure like (4) do not necessarily limit these modifiers to adjectives. The Japanese data constitute further evidence for the category-neutral nature of direct modification structure.

It may be worth stressing that the observation reported in this squib indicates that nouns should be prevented from being used as indirect modifiers. If indirect modification means being placed in a (reduced) relative clause (see footnote 5), the restriction is explained by saying that semantics goes wrong. Chuugoku-no kabin ‘Chinese vase’ ( = (10a)) is not a vase that is China, but a vase that is from China. If something like from is deleted from a relative clause, recoverability of deletion is violated.8 In the structure for direction modification in (4), on the other hand, one can say that the relevant functional head supplies the meaning of origin that is needed for proper interpretation. The same logic applies to material nouns as well. The point is moot in the case of denominal adjectives, since one may attribute the crucial semantic information to the adjective-forming suffix, but it becomes strong and transparent in the case of nouns, for which there is no other source. Thus, the structure in (4) receives support from cases where a noun directly modifies another noun.

Conversely, one may say that the semantic contribution of the functional heads in (4) blocks relative clauses, full or reduced, from direct modification, as long as relative clauses function semantically as simple one-place predicates rendered neutral with respect to the semantic distinction that characterizes direct modification. Future research should investigate the semantic properties of the functional heads posited in (4) in detail.

Notes

Thanks are due to Ken Hiraiwa and two anonymous reviewers for helpful comments. Material in this squib was presented in talks at Meiji-Gakuin University (October 2010) and Tsuda College (December 2010). I would like to thank the audiences for lively discussion. The research reported in this squib was supported by Grant-in-Aid for Scientific Research (C) 22520492 from the Japan Society for the Promotion of Science.

1 I put aside the question of whether no in cases like (10) is a genitive marker or a linker (Watanabe 2010), though it will be glossed as genitive in the text. A clear instance of the linker is illustrated in (i), where a noun is modified by a relative clause that ends with a predicative noun.

(i)

  • Taroo-ga shujinkoo-no monogatari

  • Taro-NOM protagonist-LINKER story

  • ‘a story in which Taro is the protagonist’

2 (15b) becomes acceptable even without a pause under the reading in which the contextually unique Chinese vase is small. This reading may involve focusing of the nationality modifier without a pause. Alternatively, if this reading allows merger of nationality/origin modifiers in the higher indirect modification domain, the data will be accommodated. See the discussion below of indirect modification. No complications of this sort arise for (15a). This contrast points again to the hierarchical attributive structure lying behind these modifiers.

3 Japanese lacks the kind of denominal adjectives exemplified by presidential, too. The corresponding noun shows up as part of a noun-noun compound in Japanese, as in (i).

(i)

  • daitouryou-kouho

  • president-candidate

  • ‘presidential candidate’

4Truswell (2009) observes that the classes of intersective adjectives including shape, color, nationality, and material are freely ordered in English as in (i), contrary to Scott’s (2002) claim.

(i)

  • a.

    wooden French mantel clock

  • b.

    French wooden carriage clock

Sproat and Shih (1991:588–589) make a similar observation, adding that length of adjectives is a factor that can reverse the ordering. I hasten to note that these intersective adjectives nevertheless must occur lower than subsective adjectives including size, as shown in (2). It is an important task for future research to discover where this difference between English and Japanese comes from. Sproat and Shih (1991) observe further that Mandarin allows only one adjective out of each of the larger direct modifier classes (subsective and intersective). Thus, there is more than one parametric option in the domain of direct modification.

5Cinque (2010:55) states that indirect modification adjectives occur in reduced relative clauses. It is a completely open question whether Japanese has anything like a reduced relative clause.

6 A reviewer points out that a Google search suggests that another modifier can be inserted between the material modifier and the head noun, as in stone English cottage. To the extent that nationality adjectives resist inclusion as part of a compound, this is another piece of evidence that we are dealing with phrasal configurations here.

7 Ayano also discusses furui yuuzin ‘old friend’, but the use of the adjective smacks of translation to me. To express the intended notion, a PP with a nominalized adjective as its object must be used, as in (i).

(i)

  • furuku-kara-no yuuzin

  • old.time-from-LINKER friend

8 A reviewer asks whether chuugoku-no kabin might involve a relative clause derived from the apparently predicative use of chuugoku found in cases like (i), which can mean that this vase is from China.

(i)

  • Kono kabin-wa chuugoku-da.

  • this vase-TOP China-COP

But (i) can also mean that this vase will be exported to China, depending on the context. This type of variable, context-dependent interpretation arises in so-called eel sentences, illustrated in (ii), where two nominals are connected by what appears to be the clause-final copula.

(ii)

  • Boku-wa unagi-da.

  • I-TOP eel-COP

(ii) can mean ‘I order an eel’ or ‘I don’t like eels’ or can have any other reading depending on the context. Research by Noda (2001) and Sano (2003) converges on the view that eel sentences involve ellipsis of the presupposed predicate portion (an open proposition), which gives rise to a highly context-dependent interpretation. Therefore, it is a mistake to regard eel sentences as involving a copula. The apparent copula in (i) and (ii) is actually a focus marker.

The origin reading in the modification structure, on the other hand, does not need any contextual support. Crucially, it is the only interpretation possible for cases like (15a), which suggests that the relative clause analysis that makes use of the structure underlying (ii) is ruled out.

References

References
Aljović
,
Nadira
.
2010
.
Syntactic positions of attributive adjectives
. In
Adjectives: Formal analyses in syntax and semantics
, ed. by
Patricia Cabredo
Hofherr
and
Ora
Matushansky
,
29
51
.
Amsterdam
:
John Benjamins
.
Ayano
,
Seiki
.
2010
.
Revisiting beautiful dancers and complete fools in Japanese
. In
Proceedings of the 6th Workshop on Altaic Formal Linguistics (WAFL 6)
, ed. by
Hiroki
Maezawa
and
Azusa
Yokogoshi
,
97
111
.
MIT Working Papers in Linguistics 61
.
Cambridge, MA
:
MIT, MIT Working Papers in Linguistics
.
Baker
,
Mark
.
2003a
.
Lexical categories
.
Cambridge
:
Cambridge University Press
.
Baker
,
Mark
.
2003b
.
“Verbal adjectives” as adjectives without phi-features
. In
Tokyo Conference on Psycholinguistics (TCP) 4
, ed. by
Yukio
Otsu
,
1
22
.
Tokyo
:
Hituzi Syobo
.
Cinque
,
Guglielmo
.
1994
.
On the evidence of partial N movement in the Romance DP
. In
Paths towards Universal Grammar: Studies in honor of Richard S. Kayne
, ed. by
Guglielmo
Cinque
,
Jan
Koster
,
Jean-Yves
Pollock
,
Luigi
Rizzi
, and
Raffaella
Zanuttini
,
85
110
.
Washington, DC
:
Georgetown University Press
.
Cinque
,
Guglielmo
.
2010
.
The syntax of adjectives
.
Cambridge, MA
:
MIT Press
.
Giegerich
,
Heinz J
.
2004
.
Compound or phrase? English noun-plus-noun constructions and the stress criterion
.
English Language and Linguistics
8
:
1
24
.
Kitagawa
,
Chisato
, and
Claudia N. G.
Ross
.
1982
.
Prenominal modification in Chinese and Japanese
.
Linguistic Analysis
9
:
19
53
.
Larson
,
Richard
, and
Naoko
Takahashi
.
2007
.
Order & interpretation in prenominal relative clauses
. In
Proceedings of the 2nd Workshop on Altaic Formal Linguistics (WAFL 2)
, ed. by
Meltem
Kelepir
and
Balkız
Öztürk
,
101
119
.
MIT Working Papers in Linguistics 54
.
Cambridge, MA
:
MIT, MIT Working Papers in Linguistics
.
Liberman
,
Mark
, and
Richard
Sproat
.
1992
.
The stress and structure of modified noun phrases in English
. In
Lexical matters
, ed. by
Ivan A.
Sag
and
Anna
Szabolcsi
,
131
181
.
Stanford, CA
:
CSLI Publications
.
Nishiyama
,
Kunio
.
1999
.
Adjectives and the copulas in Japanese
.
Journal of East Asian Linguistics
8
:
183
222
.
Noda
,
Hisashi
.
2001
.
Unagi-bun toiu gensou
[An illusion of eel sentences].
Kokubungaku
46
(
2
):
51
57
.
Sano
,
Tetsuya
.
2003
.
On the so-called “eel-sentences” in Japanese
. In
Empirical and theoretical investigations into language
, ed. by
Shuji
Chiba
,
604
611
.
Tokyo
:
Kaitakusha
.
Scott
,
Gary-John
.
2002
.
Stacked adjectival modification and the structure of nominal phrases
. In
Functional structure in DP and IP
, ed. by
Guglielmo
Cinque
,
91
120
.
Oxford
:
Oxford University Press
.
Sproat
,
Richard
, and
Chilin
Shih
.
1991
.
The cross-linguistic distribution of adjective ordering restrictions
. In
Interdisciplinary approaches to language: Essays in honor of S.-Y. Kuroda
, ed. by
Carol
Georgopoulos
and
Roberta
Ishihara
,
565
593
.
Dordrecht
:
Kluwer
.
Truswell
,
Robert
.
2009
.
Attributive adjectives and nominal templates
.
Linguistic Inquiry
40
:
525
533
.
Watanabe
,
Akira
.
2010
.
Notes on nominal ellipsis and the nature of no and classifiers in Japanese
.
Journal of East Asian Linguistics
19
:
61
74
.
Yamakido
,
Hiroko
.
2005
.
The nature of adjectival inflection in Japanese
. Doctoral dissertation, Stony Brook University.