Information structure is another factor that can affect the rigidity of the adjective order. If we take the DP to be analogous to the CP, then this is not surprising as the canonical order of constituents in the clause can be violated when one of the constituents is focussed or topicalised. This is demonstrated in the English example below, where the unmarked SVO order can be altered when the object bears focus.23
(67) a. Mich`ele ate the chocolate.
b. the chocolate Mich`ele ate (not the biscuits).
Rizzi (1997) argues that the left periphery of the clause consists of functional heads dedicated to information structure, such as Top0 for topic and Foc0 for focus. Any topicalised or focussed constituent will move to the functional pro-jections of these heads, hence the noncanonical ordering.
22Complement PPs of the noun do not pose a problem if we assume that these are merged higher than adjectives as in Adger 2013, or that P0 is merged high in the extended nominal projection as in Cinque 2010, following Kayne (2004, 2005).
23Focus is marked with small capitals.
Similarly, if an adjective inside the nominal domain is focussed then it can surface in a noncanonical position. While ‘old’ appears to the left of ‘Cypriot’ in the unmarked order in (68), it can appear in the reverse order if the nationality adjective is focussed.
‘He played the old Cypriot song that his grandmother had taught him’
‘We sang the cypriot old song after all, (not the Cretan one).’
Drawing upon Rizzi’s (1997) analysis of the left periphery in the clause, several linguists have argued that there is also a Focus projection in the left periphery of the DP (Giusti 1996, 2005; Dimitrova-Vulchanova and Giusti 1998; Aboh 2004;
Truswell 2004). The violation of the cross-linguistic adjective order in the example above is thus accounted for, by assuming that the focussed nationality adjective has moved above the non-focussed adjective to FocP. This is represented below:
(69) DP
The above structure shows focalisation of a phrasal adjective. If, however, direct modification adjectives can also be heads as argued in the previous section, the question that arises is whether head-modifiers can also be focussed. My claim is that only phrasal adjectives can undergo movement to a Focus position. Sup-porting evidence for this comes from the unacceptability of the examples in (70).
The non-modified nationality adjective, which as argued in the previous section modifies the noun as a head, is forbidden from appearing to the left of the phrasal adjective even when it is focussed. Movement of a head-modifier to a Focus po-sition is blocked as an X0 cannot move to a phrasal position as predicted by Emonds’ (1976) Structure-Preserving constraint, as well as the Head Movement Constraint (Travis 1984).
(70) a. *a cypriot very Cypriot man.
b. *a red very red cherry.
An alternative view is that there need not be a functional projection dedicated to focus, in order to account for the discourse-related word order variation in the clause. Instead, the noncanonical order is the result of effects in the inter-face between syntax-semantics, and syntax-information structure (Neeleman and Vermeulen 2012, and other chapters in the same book). Szendr˝oi (2012) extends this idea to the nominal phrase, and she claims that the noncanonical order of adjectives in English is triggered by scope requirements. In particular, an adjec-tive moves to mark its sister as the Domain of Contrast (DoC), as formalised by Neeleman et al. (2009):
(71) DoC Marking
The sister of a moved contrastive focal (or topical) constituent, XP, is interpreted as the domain of contrast for XP.
Szendr˝oi (2012:195) argues that adjectives which are accented and contrastively focussed do not always have to move, and she provides the example in (72) as evidence of this. Movement, according to Szendr˝oi only takes place when the adjective has to mark a domain of contrast, as represented in (73b).
(72) My friends all drive big cars, but only I drive a big black car.
(73) a. I drive a black big car.
b.
A1 black
YPDoC A2
big <A1> N car
Szendr˝oi (2012:195–196) provides the contexts in (74) to demonstrate the inter-pretational difference between (72) and (73b). The only available context for the noncanonical order black ≻ big is the first one, where the domain of contrast is the already established set of big cars. For the canonical order, however, both contexts are available. What is given in the discourse is the set of cars. The domain of contrast is not marked, and, consequently, the contrast is based on contextual clues.
(74) a. Context 1: In this car park you can see my friends’ cars and my car.
There is a bunch of big cars here. They are of many different colours.
Scenario 1: BLACK RED BLUE WHITE BLUE YELLOW b. Context 2: In this car park you can see my friends’ cars and my
car. There are cars of many different colours. Some of them are small, but there are big cars of every colour.
Scenario 2: black BLACK black black red red red RED red red
...
(75) a. I drive a black big car. Context 1: X Context 2: # b. I drive a big black car. Context 1: X Context 2: X It is not immediately obvious what the difference is between assuming a dedicated Focus position in the extended nominal projection and Szendr˝oi’s analysis, as they both involve fronting the focussed constituent. Szendr˝oi’s analysis, however, allows more flexibility in the order. This is because the focussed constituent does not move to a fixed position, but rather, it moves above the XP that needs to be marked as the DoC, which could vary from case to case. For instance, in (76a) we see that the focussed adjective moves in a position above another adjective, but lower than the numeral, as the DoC is long dresses. In (76b), on the other hand, the focussed adjective moves above the numeral since the numeral is part of the DoC six children.
(76) a. Orm’s two red long dresses are not as nice as her three black ones.
b. Carol’s horrible six children made life miserable for her second husband not as tall as her older ones.24
24Example taken from Andrews 1983:697, cited in Szendr˝oi 2012:199.
From what we have seen in this section we can conclude that information struc-ture is another contributing factor to the flexibility of adjective ordering. On the other hand, Cinque (2010:59) claims that movement of adjectives to a focus position or scope reorderings do not suffice to account for reversals of the un-marked order. Reversals of the unun-marked order are, according to Cinque, the result of merging the leftmost adjective inside a RRC. He bases this argument on the unacceptability of the reverse order in the examples in (77). If the reorderings were a mere case of focus/scope fronting, then we would expect that the following orders would be acceptable when the “displaced” adjective is focussed, but they are not. Cinque argues that the ungrammaticality is, instead, derived from the fact that the indirect source of modification is unavailable to these adjectives, as they are nonpredicative.
(77) a. *He is a nuclear young physicist (cf. a young nuclear physicist) b. *He is a heavy former drinker (cf. a former heavy drinker) c. *He is a hard alleged worker (cf. an alleged hard worker)
(Cinque 2010:59, (66)) The same argument could be extended to the examples given earlier in (24) and (25), repeated below:
(24) a. She is a tall electrical engineer.
b. *She is an electrical tall engineer.
c. *An engineer who is electrical.
(25) a. She is an alleged heavy drug-user.
b. *She is a heavy alleged drug-user.
c. #A drug-user who is heavy. or *A drug-user who is alleged.
The fact that focus fronting is unavailable in the above examples, however, can also be accounted for if we analyse these adjectives as heads. Building on what was proposed in the previous section, it appears that the adjectives in these examples are found low in the extended nominal projection, where adjectives modify the noun as heads. The assumption that these adjectives are merged in a low position is confirmed by the fact that when another adjective intervenes between them and the noun as in (77), the sentence is not ungrammatical, but we get a new reading in which the adjective loses its initial interpretation.
For instance, in heavy drinker and heavy drug-user the adjective heavy mod-ifies the event of drinking and drug-using respectively. If heavy moves above the adjective to its left, this will not result in ungrammaticality but, instead, it will give rise to the non-event readings ‘a former drinker who weighs a lot’ for (77b) and ‘an alleged drug-user who weighs a lot’ for (25b). In other words, heavy simply modifies the individual denoted by the result noun and it no longer has access to the event reading of the derived nominal. I take this interpretational distinction to be the result of the adjective modifying the noun as a head when heavy modifies the event, but as a phrase when it modifies the individual.
Consequently, the fact that the adjectives in the above examples cannot be focussed could be accounted for by the unavailability of heads to undergo focus fronting as was claimed earlier in this section. Focus fronting can, nevertheless, still account for other instances where ordering restrictions are lifted when an adjective is focussed.