Andrew Radford - Linguistics An Introduction [Second Edition]

(444) a.

Go nursery … Lucy go nursery

b.

Push Stevie … Betty push Stevie

c.

No touch … This no touch

d.

Want that … Andrew want that

e.

Plug in … Andrew plug in

Hyams went on to argue that apparently subjectless child sentences such as

those in (443) have null nominative ‘little pro’ subjects (like those found in Early Modern English, see section 22), so that a child sentence such as Want more apple would have the fuller structure indicated informally in (445):

(445)

pro want more apple

Here, the child is viewed as using the null nominative pronoun pro where an adult would use the overt nominative pronoun I. The more general conclusion which

Hyams drew was that Child English (at the relevant stage) is a null subject language –

i.e. a language which allows finite verbs to have a null pro subject. If this were so, it

352

senten ces

would provide an obvious challenge to the claim that children correctly set parameters from the outset, since adult English is not a null subject language.

However, there are reasons to be sceptical of Hyams’ claim that English

children initially mis-set the Null Subject Parameter and hence misanalyse

English as a language which allows finite verbs to have a null nominative pro subject like that found in Early Modern English (EME). We saw in our earlier discussion of EME (section 22, pp. 314ff.) that null nominative pro subjects are only licensed in EME because finite verbs raise to T and (by virtue of the rich agreement inflections they carry) can locally identify a null pro subject in spec-TP.

However, in Child English, verbs never raise to T (as we see from the fact that children never produce sentences like *Teddy likes not spaghetti in which the verb likes moves from V to T across the intervening negative particle not), and often children’s verbs carry no agreement inflection at all (e.g. they may say Teddy want ice-cream rather than Teddy wants an ice-cream). For reasons such as these, it is unlikely that children’s ‘missing’ subjects are instances of the null nominative pronoun pro found in EME.

An alternative analysis has been put forward in more recent work by Luigi

Rizzi, who argues that omission of the subject in child sentences like (443) is attributable to a separate phenomenon of truncation whereby one or more (weak or unstressed) constituents at the beginning of a sentence can be ‘silent’ (and so have a null spellout/realisation). This phenomenon of truncation is also found in colloquial adult English, e.g. in sentences such as (446):

(446) a.

Can’t find it (= I can’t find it)

b.

Know anything about it? (= Do you know anything about it?)

c.

Time is it? (= What time is it?)

d.

Nice day, isn’t it? (= It’s a nice day, isn’t it?)

As these examples illustrate, truncation affects one or more unstressed words at the very beginning of a sentence (I in 446a, do you in 446b, what in 446c and it’s a in 446d). In children’s grammars, it even seems to extend to wh-pronouns, which are sometimes omitted from questions (resulting in null operator questions). So, alongside overt operator questions like (447), a girl called Claire, aged two years to two years, one month, produced null operator questions such as those in (448) (the recordings were made by Jane Anne Collins Hill):

(447)

Where girl go? Where pencil go? Where cow go? Where the horse go? What

kitty doing? What squirrel doing? What lizard doing? What the dog doing?

What the cow say?

(448) a.

Bunnies doing? (= What are the bunnies doing?)

b.

Mommy gone? (= Where has Mommy gone?)

c.

This go? (= Where does this go?)

If children’s null operator questions like (448) are the result of truncation, a natural suggestion to make is that children’s null subject sentences like (443) are also the result of truncation (and not of a mis-setting of the Null Subject Parameter).

Children’s sentences

353

Empirical evidence in support of the truncation analysis of children’s ‘missing’

subjects comes from research done by Virginia Valian. She noted that English children omit subjects only in main clauses, never in complement clauses. If Child English were a genuine null subject language which allowed any finite verb to have a null subject, we should expect that children would omit subjects in finite complement clauses just as frequently as they omit them in finite main clauses.

But Valian’s study showed that while English children frequently omit subjects in finite main clauses, they never do so in finite complement clauses (whereas a group of young Italian children she studied frequently omitted subjects in finite complement clauses, as we would expect if they had correctly identified Italian as a null subject language). This seems to provide us with conclusive evidence that the null subjects used by English children are not the result of mis-setting the Null Subject Parameter but rather are the consequence of some independent process such as truncation. And this in turn enables us to continue to maintain the

parameter-setting model of acquisition under which children from the very outset quickly arrive at a correct setting for each parameter.

But there is a further complication which we need to take account of before we can be sure that our conclusion is correct, and this relates to the fact that children often omit subjects in wh-questions. So, for example, alongside wh-questions with overt subjects such as (447) and (448), Claire (at the same age) produced wh-questions with null subjects like (449):

(449) a.

What doing? (= What are you doing?)

b.

Where go? (= Where did it go?)

c.

What do? (= What shall I do?)

The null subject in such sentences cannot be the result of truncation, since a subject pronoun can be truncated only if it is the first word in a sentence (or if any word preceding it has itself been truncated, as with do in 446b), and it seems reasonable to assume that the wh-pronouns what/where are the first words in the sentences here, not the ‘missing’ subject pronouns you/it/I. So what precisely is the nature of the null subject in the examples in (449)?

An important clue comes from the fact that the clauses in (449) appear to be non-finite, in the sense that they contain no finite verb or auxiliary (e.g. they lack the finite auxiliaries are/did/shall which appear in their adult counterparts).

Now, we already know from our earlier discussion in section 20 that non-finite clauses in adult English (such as those bracketed below) allow a null ‘big PRO’

subject:

(450) a.

I intend [PRO going to Sri Lanka for my holidays]

b.

I intend [PRO to go to Sri Lanka for my holidays]

This suggests that the ‘missing’ subject in the non-finite wh-questions in

(449) may also be PRO, and hence that (449a), for example, has the simplified structure (451):

(451)

What PRO doing?

354

senten ces

Evidence in support of this analysis comes from the fact that English children typically don’t use null subjects in finite wh-questions – i.e. they don’t produce sentences such as the following (the asterisk here serves to indicate a non-occurring structure):

(452) a.

*What are doing? (= What are you doing?)

b.

*What did say? (= What did he say?)

c.

*Where have been? (= Where have you been?)

Why not? The answer is that children’s null subjects in wh-questions are instances of PRO, and PRO can occur only as the subject of a non-finite clause, not as the subject of a clause containing a finite auxiliary such as are/did/have) (exercise 1).

Non-finite clauses in Child English