Word Order- Classifiers and phrasal dependents in DP

Various surface orders can be observed with respect to classifiers. For example, Simpson (2005) notes the following orders among Southeast Asian languages.

Simpson argues for an antisymmetric (Kayne 1994) movement analysis; he assumes that the various elements are heads, whereas I am assuming that adjectives and numerals, at least, are phrasal dependents. Modulo these differences, a movement account can be simply characterized in the following terms: Chinese reflects something like the base order.¹ In all of the other languages, Nmoves to the left of the adjective. If relative clauses are taken to be attached to the left just above adjectives, then the Thai/Khmer pattern (henceforth Thai) and the Hmong/Malay/Vietnamese (henceforth Malay) pattern also require an additional step of movement of [N Adj] to the left of RC. In Burmese and Thai, the [RC N Adj] constituent moves across the Num– Cl sequence. In Thai and Malay, a constituent containing the numeral moves to the left of the demonstrative. The Burmese pattern is outlined in tree form below, using the labels established above; in particular, the relative clause is taken to attach above sort, and the classifier is assumed to be located in the head of unit.

In the Malay pattern, I use a convention from Koopman and Szabolcsi (2000) of superscripting a “+” to the node which includes nP and a landing site for roll-up movement, simply in order to have labels for its trace.

The Thai pattern combines both movements: the fronting of a projection of sort, as in Burmese, and the fronting of a projection of unit, as in Malay. The order is the equivalent of right-adjoining the phrasal modifiers, except that the unit classifier follows the numeral.

It is typical cross-linguistically of classifier languages that they normally do not separate the numeral and classifier (though it is not universally true; cf. Allan 1977, or example 1b above). If only maximal projections move, and the numeral is in the specifier of the classifier, then this is expected.

Note also that the position of the relative clause must be considered more carefully, and is identified here with sortP only tentatively and for purposes of illustration.

1 I have added the relative clause position to the Chinese line-up; Simpson does not discuss relative clauses in Chinese. An example, from Zhang (2004), showing the order of modifiers:

(ii) na liang ge wo tidao de nianqing ren

that two cl I mention de young person

“those two young people I mentioned”