Counting things in Japanese is tricky. For one thing, Japanese uses two kinds of numbers--a native series, and another series borrowed from Chinese.

But for another thing, Japanese numbers invariably end in what's called a numeral classifier (sometimes just called a classifier for short), which tells you something about the thing being counted; so there are classifiers like hon 'long thin thing', dai 'machine', wa 'bird', and so on. Here are some examples (here I've glossed the classifiers as CL):

(1)	a.	gakusei futa-ri
		student two CL
		'two students'
	b.	hon  ni- satsu
		book two CL
		'two books'

	c.	ninjin ni- hon
		carrot two CL
		'two carrots'

	d.	rezouko      ni- dai
		refrigerator two CL
		'two refrigerators'

	e.	kutsu ni- soku
		shoe  two CL
		'two shoes'

	f.	fune ni- sou
		boat two CL
		'two boats'
	g.	ichigo     ni- tsubu
		strawberry two CL
		'two strawberries'

	h.	haiku ni- ku
		haiku two CL
		'two haiku'

	i.	uguisu      ni- wa
		nightingale two CL
		'two nightingales'

	j.	gokiburi  ni- hiki
		cockroach two CL
		'two cockroaches'

	k.	kujira ni- tou
		whale  two CL
		'two whales'

	l.	mise  ni- ken
		store two CL
		'two stores'

The number together with its numeral classifier is called a numeral quantifier in the literature. This isn't all of the numeral classifiers in Japanese; it's just to give you an idea of what the system is like. And Japanese is very far from being the only language with numeral classifiers; we also find them in Chinese, Korean, Vietnamese, and Chol, for example.

Now, Japanese word order is pretty free. In particular, it's possible for numbers to become separated from the nouns they modify. The following sentences, for example, are all grammatical and synonymous (in these examples, NOM, ACC, and DAT are all translations for markers of what's called 'case', which we'll talk about more shortly; very roughly, case is a type of morphology that tells you what a particular NP is doing in the sentence. NOM stands for 'Nominative', which is the case that appears on subjects; ACC stands for 'Accusative', the case on objects, and DAT stands for 'dative', which appears on indirect objects):

(2)	a.	Gakusei-ga  Mary-ni  hon-o    ni-satsu ageta
		student NOM Mary DAT book ACC two CL   gave
		'The student gave Mary two books'

	b.	Gakusei-ga  hon-o    Mary-ni  ni-satsu ageta
		student NOM book ACC Mary DAT two CL   gave
		'The student gave Mary two books'

	c.	Hon-o    gakusei-ga  Mary-ni  ni-satsu ageta
		book ACC student NOM Mary DAT two CL   gave
		'The student gave Mary two books'
In all of these examples, the numeral classifier satsu tells you that the number refers to the book. And as you can see, the number and the noun being counted don't have to be anywhere near each other. If you want, you can think of the word order in (2a) as the basic one; examples like (2b) and (2c) are created by moving hon-o 'book-ACC', leaving the number that modifies the object behind where the object used to be.

Here's one fact that makes that kind of explanation attractive; although objects can freely be separated from their numeral quantifiers, subjects can't be:

(3)	a. 	Gakusei-ga  futa-ri Mary-ni  hon-o    ageta
		student NOM two CL  Mary DAT book ACC gave
		'Two students gave Mary books'

	b.    *	Gakusei-ga  Mary-ni  futa-ri hon-o    ageta
		student NOM Mary DAT two CL  book ACC gave
		'Two students gave Mary books'

	c.    *	Gakusei-ga  Mary-ni  hon-o    futa-ri ageta
		student NOM Mary DAT book ACC two CL  gave
		'Two students gave Mary books'
Here the number modifies the subject (you can tell, because it takes the classifier for human beings), and as you can see, it's not as easy to separate it from its NP as it was in the examples in (2). If we posit movement, we can account for the data in (2-3) by giving Japanese a basic SOV order, and saying that NPs can leave their numbers behind when they move around. So in (2), the object is moving towards the front of the sentence, and leaving its numeral quantifier behind, right before the verb. But in (3), the subject is at the beginning of the sentence where it belongs, and the numeral quantifier has no way of ending up separate from it.

We'll be seeing more evidence for looking at Japanese that way eventually. For now, whether you buy that story or not, here's a summary of the data in (2-3); objects can be freely separated from their numeral quantifiers, and subjects cannot be.

Examples like (4) look like bad news for that generalization:

(4)	Gakusei-ga  ano  otoko-ni futa-ri korosareta
	student NOM that man DAT  two CL  were-killed
	'Two students were killed by that man'
On an approach to the passive that posits movement, however, we have a natural account of why subjects of passives can leave numeral quantifiers behind. In Japanese, as in English, the subject of a passive begins in object position, and moves to the specifier of TP--and in Japanese, when it does so, it has the option of leaving its numeral quantifier behind:


Back to A-movement.