Table 5. Frames produced by fewer than 5 children and order of emergence

Learning the English auxiliary 



by 4 of the children and its mean rank is very close to the lowest rank in Table 4. Note

that there are no tensed auxiliaries or wh-frames in Table 4.

The one child who shows some clear variation from this overall pattern is Ivan. In

addition to producing almost all the frames in Table 4, he also has high ranks for 8 of

the structures in Table 5. Six of these 8 frames are yes/no-questions and the high rank

for Ivan’s Are you –ing? explains why its mean rank for the group of children is so high.

Of course, this is partly because he started so late that there were fewer sessions before

the end of the study that could acquire lower ranks but it is also because he showed

both rapid development overall and a unique development of yes/no-questions at the

same time that he was developing the same group of frames as the other children.

For some children and some auxiliaries, a range of frames sharing a particular

auxiliary also share the same rank (i.e. become productive in the same month). This is

the case, for instance, for the 3 CAN frames for Kathy and Les. However if we look at

the ranks within the 5 auxiliary groups for each of the 6 children, we can see that this is

only the case for 6 out of the 30 possibilities, again suggesting that many of these frames

are being learned separately and not initially as part of a more general system. This is

despite the fact that all three main auxiliaries together with two modals are represented

in the frames in Table 4. There is also a yes/no-question (Can X -?) and a number of

negated frames. The implication is that the mere presence of large numbers of utterances containing a wide range of auxiliaries and syntactic structures does not guarantee

the existence of a fully abstract auxiliary syntax in these children’s grammars.

2.4.3 Evidence for developing schematicity and generalization

Subjects and verbs

In a usage-based account, constructions can be productive while still being only partially schematic. Many of the children’s early auxiliary frames are indeed highly productive with a wide range of verbs being used. There are a number of possible indications of increasing schematicity of these constructions and of low-scope constructions

building up towards a more auxiliary-wide syntax. For instance, increasing schematicity can be seen in the development of the subject slot in the frames in Table 4. Those

frames with the highest ranks almost all come in initially either without a subject or

with a fixed subject, usually I. Frames with lower ranks that come in later, together

with those in Table 5, are more likely to show subject variation in the three utterances

that go up to make the frame: 54% of the frames in Table 4 come in with variable subjects as opposed to 67% in Table 5. Thus in using these early low-scope frames, children are showing a developing knowledge of the range of subjects and of verbs that can

be placed in and across the frames.

Errors

Errors can also be informative as to what is formulaic and which parts of the system

may be developing. Errors were noted if they concerned the subject, auxiliary form,

main verb form or tag. All the children’s non-tag question errors are listed in Table 6,







Elena Lieven



though I will only discuss those that show some systematicity. Errors in the first utterances contributing to a frame are also indicated in Tables 4 and 5. Although the relative

thinness of the sampling makes it difficult to draw hard and fast conclusions about

how systematic these errors are, we will briefly consider them as possible indications

for developing generalizations and for possible links between different auxiliaries. First

we should note that, of the 62 frames in Table 4, only two have errors in the utterances

that contribute to the frame being classed as productive.10 On the other hand, 20 of the

57 frames in Table 5 have some kind of error in their initial three utterances and these

errors are divided almost equally between question frames (37.5% errors) and nonquestion frames (36% errors). This difference suggests that, as the later frames were

developing, the children may indeed have been starting to grasp some of the more

general aspects of the auxiliary system and be groping for the correct form for both the

auxiliary and for the main verb while having available a variety of more or less entrenched lexically-specific constructions with auxiliaries. We can look at this in a bit

more detail by consulting Table 6.

I have divided the errors in Table 6 into five types together with a miscellaneous

column that includes a number of wh-inversion errors. The first two columns show

errors in the form of the main verb following auxiliaries BE and HAVE respectively

and the third shows errors where the children appear to have used BE for HAVE or

vice versa. The fourth column lists errors with DO including the wrong form of the

main verb, some ‘over-tensing’ errors and some possibly emphatic uses of DO. The

fifth lists all agreement errors.

One possible way to think about the errors in Table 6 is in terms of what they tell

us about (a) the open slot in the construction and (b) the child’s knowledge of the auxiliary system beyond any particular lexically-specific construction. Thus April, Les and

Mavis’ errors with HAVE in column 2 involve using an incorrect past participle, often

of an irregular verb, blowed for blown, flied for flown, drank for drunk, maked for made

(as noted by Pinker 1984) which suggests that the slot in this frame is already linked to

some meaning associated with immediate past. On the other hand, the errors in the

verbs that Charles uses with this construction are less clearly marked for past which

may mean that he is still groping for the correct form-function mapping for this slot. A

further example is the ‘double tense-marking’ errors of April and Ivan with did (column 4). These were also noted by Pinker (1984). Again the use of past tense on the

main verb suggests that the children have identified the meaning of the slot with past

and probably also the frame. Here, though, they have not coordinated between the two,

presumably because a more abstract ability to manipulate finiteness is still developing.



10. Charles: Peter’s broke my gun. Ivan: Can’t get him out, in’t he?



2;6

– My Daddy’s do

that

– He’s not go fast



2;10

– Brian is go on

– I’m go be Joe Lean



Charles



Wrong main verb

form with is



April



Child



2;8

– Haven’t read

(present tense)

it before

2;9

– I haven’t march

anymore



2;7

– Peter’s broke my

gun



2;10

– He’s blowed

that candle out

– He’s flied away



Wrong main verb

form with have



Table 6.  The Children’s Non-Tag Question Errors



2;11

– Mine is dropped

off



Is for have or vice

versa



2;11

– It doesn’t

crossen that

– Where this one

goes, this goes

in here and does

this one goes in

here?



2;11

– I did found a

mouse

– I did found a

crab



Non-agreement

errors with do



2;10

– He’s haven’t got

a digger

– We’s having

dinner



2;8

– F.Xmas haven’t

got a key



Agreement errors



2;9

– What Brian is

going?



2;7

– Why that car

wouldn’t go?

2;10

– Them haven’t

made a big one



Other





Learning the English auxiliary 



Kathy



Ivan



Child



2;6

– I’m colour



2;11

– Are you make

some more

blutack?



Wrong main verb

form with is



3;0

– Have you eat

yours?



Wrong main verb

form with have

2;9

– Are you got

another car?

– Are you mended

that?

2;11

– What are you

made?

– What is you

made?



Is for have or vice

versa



2;10

– Has you got that

much?



2;11

– D’you bought it?

– Did you broke

it?

– I did fall out

2;11

– Had (for have)

you got two

cars?

– Has you got

lipstick on?

– Has you got

windows in

your car?



2;9

– Has you got

another car?



Agreement errors



2;9

– Don’t got a

hanky



Non-agreement

errors with do



2;11

– Where’s he can’t

get?



2;10

– This one’s need

sharpen



Other





Elena Lieven