Andrew Radford - Linguistics An Introduction [Second Edition]

(498)

Reconstruction of target

Realisation

a.

Ich fahre auch

ein Auto

is auch

ein auto

fahr

I

drive also

a

car

I

also

a

car

drive

‘I also drive a car’

b.

Einen Sitz

brauche ich

ein titz is brauch

A

seat

need

I

a

seat I

need

‘I need a seat’

The speech therapists’ view has been confirmed by several empirical studies, where it has been found that between 60 per cent and 70 per cent of the main

384

senten ces

clauses produced by German SLI children have the verb in clause-final position.

Similar results have been obtained in sentence-imitation tasks; when SLI children have been asked to imitate German sentences such as the targets in (498), in 60 per cent of cases they have changed the given word orders to patterns with the verb appearing at the end of the clause.

Speech therapists have taken the frequent use of verb-final patterns by SLI

children as an indication of a severe word-order deficit and have developed

sentence-pattern drills and other therapeutic techniques for teaching the children

‘proper’ German word order. However, this therapy has turned out to be unsuccessful, suggesting that the supposed word-order deficit is resistent to therapy.

At this point, a linguistic perspective can help to resolve the issue and may in fact contribute to specifying appropriate therapeutic goals for children

suffering from SLI. A syntactic analysis of the verb-final patterns German

SLI children produce shows that their sentences are not in fact as deviant as might be thought at first sight. Verb-final patterns are in fact possible in German main clauses, but only for non-finite verbs; see the discussion of German word order in section 22. For illustration, consider the example in (499). Note that verbs can, in principle, appear in two different positions in German main clauses: finite verbs must appear in the second structural position, such as, for example, hat in (499), whereas non-finite verbal elements (i.e. infinitives or participles) appear in final position, like angestellt in (499). In syntactic terms, we say that VP and TP in German are head-final, whereas the functional projection that

hosts the finite verb (i.e. CP) is head-initial (cf. the tree diagram in 500 and

section 22):

(499)

Der Adrian hat das Radio angestellt

The Adrian has the

radio

on-turned

‘Adrian has turned on the radio’

(500)

CP

DP

C'

Der Adrian

C

TP

hat

DP

T'

der Adrian

VP

T

hat

DP

V

das Radio

angestellt

In terms of this analysis, it might be suggested that SLI children have in fact acquired the correct word-order system, that is, they know that VP in German is

Syntactic disorders

385

head-final. Recall that SLI children typically produce non-finite verb forms, like infinitives or simple verbal stems, as in the realisations in (498a, b), and that these appear in clause-final position. The few finite verb forms they produce are

correctly placed in second position. Note, for example, that the same child who produced the verb-final patterns in (498) also produced sentences such as (501)

with the finite modal auxiliary will ‘want’ in second position and the infinitive haben ‘to have’ in clause-final position, the correct pattern for German:

(501)

ich

will

auto

haben

I

want

car

have

‘I want to have the car’

Thus, it seems that with respect to word order, the grammar of SLI subjects is in fact identical to that of unimpaired speakers, as all the verbs they use appear in the correct positions. The only difference between SLI subjects and normal children is that SLI children do not produce as many finite verb forms as the language requires. This is why sentence-pattern drills aimed at teaching SLI subjects verb-second patterns, in which the finite verb has moved to C, fail to show any effect: they simply miss the point. A more sensible goal for therapy would be to help the SLI subjects overcome their problems with finite verb

formation.

We conclude that the grammatical problems of SLI subjects lie mainly with

inflection, and that word order is in fact unimpaired. Within the area of

inflection, subject–verb agreement, case-marking, gender and auxiliaries appear to be more strongly affected than, for example, noun plurals (exercises 4

and 5).

In this section, we have looked at different language disorders from a syntactic perspective. The phenomenon of agrammatism is perhaps the clearest case of an impairment to the central cognitive system that underlies the production and comprehension of sentences. We saw that agrammatism affects both sentence

production and comprehension, and that the deficit can be characterised in

syntactic terms, namely as an impairment to the internal feature specification of functional categories. The phenomenon of paragrammatism, by contrast, does not seem to involve a genuine syntactic deficit. Rather, the paragrammatic errors such as blends, constituent substitutions, etc. that Wernicke’s aphasics typically produce result from a lexical disorder, specifically from word-finding problems. We also saw that in SLI subjects, the normal development of grammar is selectively impaired, and that the impairment mainly affects inflection. Word order, on the other hand, appears to develop normally in SLI subjects. The importance of the properties of functional categories and inflection which has emerged in this discussion is, of course, reminiscent of what we saw in our discussion of the syntax of normal children. The view that these aspects of linguistic structure hold the key to the essential nature of language and the human language faculty is one which is informing a great deal of current work, and we fully expect this to continue to be the case for the foreseeable future.

386

senten ces

Exercises

1.

Grodzinsky (1990) proposed a syntactic theory of agrammatism according to which the phrase-structure representations of these patients lack

syntactic features. Discuss this claim in the light of the findings below

from an elicitation experiment in which patients with agrammatism

were asked to complete sentences testing for tense and subject–verb

agreement marking. The following table shows percentages of incorrect

responses in these two conditions for simple verbs and for auxiliaries.

Agreement

Tense

Verb

3.2%

38%

Auxiliary

0%

70%

What do these results indicate?

2.

Friedmann (2001) elicited wh-questions from agrammatic Broca’s aphasics. Overall, the patients produced only 23 per cent correct

wh-questions involving wh-movement. The most typical response

was an inappropriate yes–no question which was just marked by

intonation, e.g. You have hammer? intead of Which hammer do you

like to have? Explain how this finding can be accounted for in terms of

a syntactic feature deficit account.

3.

Compare agrammatism and paragrammatism using the following data

in which two patients are trying to describe a picture illustrating

various household dangers. (Note: Pauses are indicated by dots.)

(a) You never do that with a place there, you push it and do that …

That is the same thing underneath; there’s a little one to that as

well. That you don’t have to do either. I don’t know what’s hap-

pened to that, but it’s taken that out. That is mm there without

doing it, the things that are being done.