Content uploaded by Jan de Jong
Author content
All content in this area was uploaded by Jan de Jong on Jan 30, 2017
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=ilog19
Download by: [Universitetsbiblioteket i Bergen] Date: 30 January 2017, At: 01:08
Scandinavian Journal of Logopedics and Phoniatrics
ISSN: 0803-5032 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/ilog19
Auxiliary verbs in Dutch SLI children
Gerard W. Bol & Jan de Jong
To cite this article: Gerard W. Bol & Jan de Jong (1992) Auxiliary verbs in Dutch SLI
children, Scandinavian Journal of Logopedics and Phoniatrics, 17:1, 17-21, DOI:
10.3109/14015439209099177
To link to this article: http://dx.doi.org/10.3109/14015439209099177
Published online: 11 Jul 2009.
Submit your article to this journal
Article views: 21
View related articles
Citing articles: 1 View citing articles
Scand
J
Log
Phon
17:
17-21,
1992
Auxiliary verbs in Dutch
SLI
children
Gerard
W.
Bol
and Jan de
Jong
Bol.
G.
W.,
de Jong,
J.
(1992). Auxiliary verbs in Dutch SLI children. Scand
J
Log
Phon 17: 17-21.
Symptomatology in
SLI
children has been characterised for English speaking children
in a number
of
studies. Symptoms
for
Dutch SLI children have not been described
to
the same extent.
In
this article we wanted to examine in detail a category which has
regularly been highlighted as deficient in English speaking children, namely auxilia-
ries. Available data allows
us
to examine the distribution
of
members
of
this category
in
a group
of
16
Dutch SLI children and
16
MLU matched normals. The analysis
of
our
data was based
on
hypotheses derived
from
a comparative view
of
the English
and Dutch auxiliary system. The data used are from a corpus formerly gathered by
Bol
and Kuiken
(1988).
Our
conclusion is that Dutch SLI children show
no
group
differences in production
of
auxiliaries in comparison to normals. Six out
of
16
SLI
children omitted auxiliaries.’
Key
words: specific language impairment, auxiliaries.
Jan
de
Jong,
Sarphatipnrk
30,
1072
PB
Amsterdam, the Netherlandy.
Introduction
The literature about the linguistic symptoms
of
specific
language impairment is predominantly oriented towards
English speaking children. This poses a question: what
is the relevance
of
these symptoms for other language
communities? In this article we want to take up one
problem area often mentioned for English SLI children
and assess its importance for their Dutch counterparts.
Categories which seem to present language-impaired
children with difficulties are bound morphemes and free
morphemes within closed classes, such as articles, cop-
ulas and auxiliaries (Leonard
et
al.,
1987; Rice, 1991).
The one aspect we want to address is the use
of
auxilia-
ries.
Fletcher and Peters, in their 1984 article, tried to find
categories that differentiate SLI children from normals.
They found, in a discriminant function analysis, that
one of the two categories which most reliably distin-
guished SLI children and normals was UVF, ‘Unmarked
Verb Form’. The SLI group were more likely than the
normal group
to
provide, in a clause, a verb which was a
bare item
-
uninflected for tense, and not premodified
by an auxiliary.
Johnston and Kamhi (1984:73) noticed that “error
rates in the negatives, main verb, and interrogative
1.
The analyses reported here were made by a group
of
stu-
dents at the University
of
Groningen, Department
of
Lin-
guistics. We have benefited from comments
on
previous
versions
by
Harald Clahsen, Paul Fletcher, Kristina Hans-
son
and Anne Mills.
reversal categories in large part reflected difficulties
with the auxiliary system
(..)
forms which seemed par-
ticularly vulnerable to error included the auxiliary and
infinitive markers.
”
Clahsen (1989:
904)
found, in a group
of
German SLI
children that
“all
the children use simple verbs, prefix
verbs, and modals
(..).
Auxiliaries and copulas, how-
ever, are used in only very few cases. Moreover, the
longitudinal data show that the proportion
of
deleted
elements decreases over time.
No
progress is made,
however, in auxiliaries and copulas”. Clahsen found
high deletion rates and no systematic changes when the
children get older.
In addressing the Dutch data, then, we need to con-
sider first the differences of the Dutch auxiliary system
from the English auxiliary system, to see if the pre-
dictions for Dutch might be different. In the second
place we need to
look
at the phonetic environment of
Dutch auxiliaries. Leonard
et
al.
(1987: 234) state that
“SLI children’s use
of
the copula, auxiliary and verb
inflections lags behind expectations based
on
mean ut-
terance length” and add that “contextual analyses of
copula and auxiliary use reveal clear phonetic envi-
ronment influences”. The third point
to
consider is the
data available on Dutch SLI children.
The comparison with English may seem arbitrary and
uncalled for, but it illustrates the problem of transfer-
ring foreign data, as well as the relative lack
of
in-
formation
on
linguistic symptoms of SLI in Dutch.
2
17
G.
W.
Bol
and
J.
de
Jong
Auxiliaries in English and Dutch
The systems compared
In order to compare the English auxiliary system with
the Dutch one, we take the description
of
the English
auxiliary in Radford’s 1988 textbook on transforma-
tional grammar. On page
153
he gives these character-
Auxiliaries can undergo inversion in direct ques-
tions, whereas Nonauxiliary verbs cannot (and
require do-periphrasis)
Auxiliaries can occur in tags, whereas Nonauxilia-
ry verbs cannot (and require do-tags)
Auxiliary verbs can be negated by not/n’t,
whereas Nonauxiliary verbs cannot, and require
do-negatives
Only Modals take
a
‘bare’ infinitive VP comple-
ment
Unlike other Verbs, Modals have no infinitive
form, and hence cannot be used after the infin-
itive particle ‘to’,
or
after another Modal
Unlike other Verbs, Modals have no ‘-n’ parti-
ciple form, and hence cannot be used after perfec-
tive ‘have’
Unlike other Verbs, Modals have no ‘-ing’ form.”
Of
these, only (d) applies to Dutch.
Palmer (1974) makes
a
distinction between primary
auxiliaries and modals. Modals have no flexion mor-
pheme, while primary auxiliaries do have one. This
difference also exists in Dutch. However, many
of
the
difficult aspects of the English system, as shown in
Radford
(1988),
do not seem relevant to Dutch, a lan-
guage with little use for tags, where main verbs can
stand perfectly well on their own in questions
or
nega-
tive sentences. This
is
a first indication that the
SLI
children’s problems with auxiliary use in English might
be
to
a large extent language specific. Moreover, many
aspects
of
language related
to
auxiliary problems from
the literature on
SLI
are not valid for Dutch; in Dutch
potential problems would arise in a different syntactic
context.
An
element
of
Dutch which differentiates it from
English is the use
of
independent ‘pseudo-modals’.
These have no verbal complement, but take a direct
object. Some examples:
(la) Ik kan dat
(1 b)
I
k kan dat doen
(2a) Ik moet die
(2h) Ik moet die hebben
(I
can that)
(I
can
do
that)
(I
must that one)
(I
must have that one)
If we take the (a) examples
to
be elliptical, it is impor-
tant
to
stress
that it is often not possible
to
determine
which verbal lexical item has been elided.
The phonetic environment
Leonard has directed our attention on
a
number
of
occasions
to
the influence
of
phonetic saliency on pro-
duction of bound and free closed-class morphemes.
For
auxiliaries in Dutch only one phenomenon seems rele-
vant. The Dutch auxiliary
zijn
(to be) can be produced
in a contracted form in the third person singular, though
only after a demonstrative pronoun. For example:
(3a) da’s gevallen
next
to
the full form:
(3b) dat
is
gevallen
(that has fallen)
(that has fallen)
Auxiliaries in Dutch children
In 1988,
Bol
and Kuiken presented a Dutch adaptation
of the LARSP procedure of Crystal, Fletcher and
Gar-
man
(1976).
In their research project
Bol
and Kuiken
sampled
100
utterances
of
spontaneous language for
each child, analysed them morphosyntactically and then
compared scores of groups of children, including an SLl
group, with their normals. Comparing a group
of
12
normal children, aged
3;64;0,
with a
group
of
18
SLI
children, they found no significant difference in auxilia-
ry use.
It
has
to
be pointed out, however, that their
category included all auxiliaries with a verbal comple-
ment, modals as well as others.
In a Dutch paper by Jansen
(1987)
the appearance
of
auxiliaries was explored in detail for normally devel-
oping children between two and four years
of
age. The
children belong to the corpus
of
the
Bol
and Kuiken
research. Jansen made an inventory
of
the auxiliaries
used in 48 samples, divided into four stages, conforming
to
stages
I11
to
VI
of
the Dutch profile chart.
Her
overall conclusion was that Dutch children acquire aux-
iliaries as a system, and not ‘piecemeal’. From stage
I11
(age 2;0-2;6) not
just
one auxiliary was produced, but a
set of them (see Table
1).
Table
1.
Auxiliaries
most
frequently
used
by
normal Dutch
children, Jansen
1987.
Stage Age Modals Aspect
111
2;@-2;6
kunnen gaan
(can) (to
Po)
moeten
(must, to have to)
IV
2;6-3;0
hebben
V
3;U-3;6
mogen
zijn
(to have)
(to
be
(to be)
allowed to)
(to want to)
VI
3;u;O
willen
18
Auxiliary verbs in Dutch
SLI
children
Table
2.
Mann-Whitney U tests SLI and normal children
(p<
.OS).
Variables’
Tense (T)
Aspect (A)
Modality
(M)
Passive (P)
Causative
(C)
Other
‘to
have’ (A)
‘to
bc’
(A)
inchoative (A)
‘to
go’ (A)
‘to
come’ (A)
shall/will (A)
progressive
(A)
past
part.
(A)
to
be’ (A)
‘to
remain’ (A)
capability (M)
desire (M)
Obligdtion (M)
permission
(M)
causative
(C)
utt.
with verb
utt.
with aux
Yo
aux
in
utt.
with verb
‘to do’ (C)
Mean rank Corr.
for ties
SLI
Normal 2-tailed
Signif.
5.0 4.0 .51
16.6 15.4 .72
15.9 16.0 .98
2.0 4.0
.ox
2.0
1
.o
2.0
7.8
8.3
4.7 4.2
14.4 14.7
14.4 14.5
2.5 2.5
1.5
1.5
4.0 4.0
10.1 7.8
2.0 3.0
12.0
8.6
3.4 5.6
11.9 16.6
5.5 5.5
1.5
15.6
17.4
16.5 16.8
17.5 15.5
.32
.80
.73
.93
.Y8
1
.oo
1
.oo
1
.0o
.31
.32
.I7
.I6
.12
1
.o0
.5n
.98
.55
*
According
to
Jansen, 1987.
Hypotheses
We are now
in
a position
to
formulate a strong hypothe-
sis: since the Dutch system is far less complex than the
English system, Dutch SLI children will show no prob-
lems with auxiliaries. The assumption behind this hy-
pothesis is that the reason for problems with auxiliaries
in
English is their involvement in complex syntax and
their sometimes low phonetic substance:
A
less radical formulation would be that the problems
for Dutch children exist, but are less severe than in
English. That hypothesis would leave intact the notion
that auxiliaries per se involve a certain complexity, be it
grammatical, semantic
or
otherwise.
Method
From the children examined in the
Bol
and Kuiken
project, we selected
16
samples
of
SLI children and
16
of MLU-matched normals. Of every child
100
analys-
able utterances were available. In every sample the
auxiliaries had already been identified and counted. We
assigned them to classes as well as to lexical items,
according to the classification Jansen used. Further-
more, omissions were counted. This is, of course, only
possible for aspect auxiliaries; as explained earlier the
omission
of
a modal cannot be reliably verified. Scores
of
both groups were compared using the Mann-Whitney
U
test.
Results
As
can be seen
in
Table
2,
none
of
the categories
showed significant differences between the
SLI
group
and the normal group.
As
for
omitted auxiliaries,
6
children showed
in
total
10 omissions in
42
obligatory contexts. The children
omitted only aspect auxiliaries: ‘hebben’ (to have)
7
times and ‘zijn’ (to be)
3
times.
For
example:
Hessel,
age
6;l
MLU
2.1:
(4)
broertje daanl
(5)
broer edaan/
(little brother <has> done
<that>)
(brother <has> done
<that>)
Joep,
age
6;2
MLU
3.5:
(6)
dat Zeppelin daanl (Zeppelin <has> done that)
(7)
ja, dat Zeppelin
gedaanl (yes, Zeppelin <has> done
that)
The correct participle in Dutch has
a
prefix ‘ge-’. The
differences in the utterances quoted are
in
the real-
isation
of
thc prefix. The right form is ‘gedaan’. Note
that this omitted auxiliary
is
not
of
the contractible
kind. The auxiliary would be ‘heeft’.
With reference to Leonard
et
af.
(1987) we looked at
the correlation between MLU and the number of auxil-
iaries, both in the normal and the SLI group. For both
groups there was a clear correlation between increase of
MLU and number of auxiliaries (see Table
3).
Discussion
These results support the strong version of
our
hy-
pothises, i.e. that, since the English system is more
complex than the Dutch,
SLI
children learning Dutch
will show
no
problems with auxiliaries. We have seen
that
no
type
of
auxiliary was used less frequently by
our
SLI
group than by the controls. Six SLI children omit-
ted auxiliaries.
When comparing
our
results with the results of Clah-
Table 3. Correlations with MLU.
Number
of
Number
of
utterances utterances with
with verbs auxiliaries
SLI
MLU
.64*
.68*
Normals MLU .61* .78”*
*
=
a
<
.01
**
=
a
c
.001
2’
19
G.
W.
Bol
and
J.
de
Jong
sen’s research on German
SLI
children, we find that
there is some discrepancy. Since German is closer to
Dutch than to English, in
so
far as the auxiliary system
is less complex, German auxiliaries should not be a
problem either. The difference between the results of
Clahsen’s and
our
studies only concerns omissions.
Of
Clahsen’s
10
SLI
children,
8
subjects omitted auxilia-
ria2.
In interpreting the differences one might look at the
MLU
range of the
SLI
children in both studies. In
Clahsen’s work they range from 1.46 to 2.84 (with one
sample reaching 3.25). The children in our study range
from 2.1
to
4.7. Only four of our
SLI
children overlap
with the
MLU
range of Clahsen’s subjects.
A
simple
prediction might be that Clahsen’s children may still
‘outgrow’ their problems, since, as we saw before, in
our
children increase in
MLU
correlated with increase
of
auxiliaries. This line
of
thinking is not fully supported
by our own data: the Dutch children who omitted auxil-
iaries had
MLUs
from 2.1 up
to
4.2.
Another way
of
looking at it might be to consider
sample
size.
Clahsen’s samples range from 96 to 217
analysable utterances, while ours are restricted to
100
utterances. The sample size being smaller, the Dutch
children may have had fewer opportunities (obligatory
contexts)
to
use auxiliaries.
Conclusion
Dutch
SLI
children showed no group differences in
production
of
auxiliaries in comparison to normals. Six
out
of
16
SLI
children omitted auxiliaries.
We assumed that auxiliaries would provide an exam-
ple
of
a functional category where English
SLI
children
face problems. Comparison
of
characteristics
of
the En-
glish and the Dutch auxiliary system reveals differences.
Without pretending to fully cover the relevant factors, it
seems that the differences between linguistic systems
go
some way towards explaining the differences in beha-
viour between
our
subjects and the children described
in the English language literature.
References
Bol.
G.
W.,
Kuiken, F. (1988).
Crammaticale analyse van
taalontwikkelingsstoornissen,
PhD dissertation, University
of
Amsterdam.
Clahsen.
H.
(1989). The grammatical characterization of de-
vclopmental dysphasia. Linguistics 27: 897-920.
Clahsen,
H.
(1991).
Child
language and developmental dys-
phasia.
Amsterdam: John Benjamins.
Crystal,
D.,
Fletcher, P., Garman, M. (1976).
Thegrammatical
analysis
of
language di.sability.
London: Arnold.
2. A paper by Hansson (1992) presented results similar to
those
of
Clahsen. In a group
of
5
children 47.5%
of
all
auxiliaries were left out, with two children omitting virtually
all auxiliaries.
Fletcher, P., Peters,
J.
(1984). Characterising language impair-
ment in children: an exploratory study. Language Testing
1:
33-49.
Hansson,
K.
(1992). Swedish verb morphology and problems
with its acquisition in languagc disordered children. Paper
presented at the 1992 Lund Symposium on child language
disorders,
2-5
May.
Jansen,
E
(1987).
De
ontwikkeling van hulpwerkwoorden in
de taal van twee- tot vierjarigc kinderen. M. A. thesis,
University
of
Amsterdam.
Johnson,
J.,
Kamhi, A. (1984). The same can be less: syntactic
and semantic aspects
of
the utterances
of
languageimpaired
children. Merrill-Palmer Quarterly
30:
65-86.
Leonard,
L.
B.,
Sabbadini, L., Leonard,
J.
S.,
Volterra, V.
(1987). Specific language impairment in children: a cross-
linguistic study. Brain Lang 32: 23S252.
Palmer,
F.
(1974).
The
Engiish
verb.
London: Longman.
Radford, A. (1988).
Transformational grammar.
Cambridge:
Cambridge University Press.
Rice, M. (1991). Children with specific language impairment:
towards a model
of
teachability. In:
E.
Krasneger (Ed.).
Biobehavioral foundations
of
language.
Hillsdale,
N.J.
:
Erl-
baum.
Sammanfattning
Hjalpverb
hos
hollandska sprikstorda barn
Artikelforfattarna inleder med att konstatera att littera-
turen om sprikliga symtom hos barn med sprikstorning
huvudsakligen handlar om engelsksprikiga barn.
De
frigar sig darfor: ar symtomen relevanta ocksi for an-
dra sprik?
Problem med hjalpverbssystemet och verbandelser
har ofta namnts i litteraturen som nigot som skiljer
sprikstorda barn frin barn med normal sprikutveck-
ling. Forfattarna har darfor valt att jamfora sprikstorda
hollandska barns beharskning av hjalpverbssystemet
med vad som rapporterats om sprikstorda engelska
barn.
Det engelska hjalpverbssystemet ar jamfort med det
hollandska, komplext och
de
flesta av sdrigheterna,
t
ex
att hjalpverb kan genomgi inversion
i
direkta frigor
och kan negeras med “not”/”n’t”, medan icke-hjilpverb
har kraver do-omskrivning, att modala hjalpverb ej har
infinitiv, particip- eller -ingform, ar sprikspecifika.
De hypoteser forfattarna staller upp ar: a) eftersom
det hollandska hjalpverbssystemet ar enklare, bor det
inte villa problem for sprikstorda hollandska barn
(stark hypotes); och
b)
hollandska sprikstorda barn har
problem med hjalpverb, men lindrigare an engelska
barn (svagare hypotes).
1
studien, som ornfattar
16
barn med sprikstorning
och 16 sprikligt matchade kontroller frin vilka
100
yt-
tranden vardera analyserats, fann man inga skillnader
vad betraffar utelamnande av hjalpverb (“ha” och
“vara”). Dvs hjalpverben villar inte storre problem for
sprikstorda an for sprikligt nomalutvecklade barn.
Man fann daremot ett klart samband mellan okande
MLU
(yttrandemedellangd) och okande anvandning av
hjalpverb.
Forfattarna menar att resultaten stoder den starka
20
Auxiliary verbs
in
Dutch
SLI
children
hypotesen: det engelska hjalpverbssystemets komplexi-
tet giir det svirare for engelska sprikstorda barn att
tillagna sig, medan det enkIare hollandska systemet inte
villar storre problem.
Forfattarna noterar en skillnad gentemot Clahsens
resultat frin tyska barn. Eftersom tyskans hjalpverbs-
system Br mera likt hollandskans an engelskans, borde
intc heller tyska barn ha problem med hjalpverb. Detta
visar sig dock vara fallet
i
Clahsens studie, dar de flesta
sprHkstorda barnen utelamnar hjalpverb. Skillnaden
skulle kunna forklaras av att de flesta barnen i den tyska
studien har Iagre
MLU
an de som deltog
i
den hol-
Iandska studien.
Forfattarnas slutsats ar att sprikstorda hollandska
barn har mindre problem med hjalpverb an sprikstorda
engelska barn och att detta skulle kunna forklaras av
skillnader
i
komplexitet mellan de bida sprbken. Resul-
taten och slutsaten visar hur viktigt det
Cr
att gora tvar-
sprikliga studier av sprikliga symtom vid sprikstor-
ning.
Y
h
teenve
to
Hollantilaisten kielellisessa
kehtyksessa
viivistyneiden lasten apuverbit
Kirjoittajat toteavat, etta kielellisia oireita koskettava
kirjallisuus kasittelee paaasiasssa englantia puhuvia lap-
sia. Ovatko oirekuvaukset mielekkaita myos muissa
kielissa?
Apuverbien ja verbipaatteiden hallinnan on havaittu
erottelevan kielenkehityksen hairioista karsivia lapsia
normaalisti kehyttyneista. Tutkijat vertasivat hollanti-
laisten lasten apuverbijarjestelman hallintaa englannin-
kielesta saatuihin tuloksiin.
Englantilainen apuverbijarjestelma on huomattavasti
monimutkaisempi kuin hollantilainen ja useimmat vir-
heet ovat kielisidonnaisia.
Tutkimushypoteesit olivat: a) koska hollantilainen
jarjestelma on yksinkertaisempi, eikii se sitten aiheuta
ongelmia kielellisessa kehityksessa viivastyneille lap-
sille? ja
b)
hollantilaisilla kielihaririolapsilla on ongel-
mia apuverbeissa, mutta lievempia kuin englantilaisilla.
Tutkimusryhmassa ja vakioidussa verrokkiryhmassa
oh
16
lasta kummassakin. Tutkittiin sata ilmausta eika
havaittu eroja apuverbien poisjattamisessa.
Eli
apuver-
bit eivat sinansa aiheuttaneet ongelmia tutkimusryhman
lapsille. Sen sijaan havaittiin kasvavan ilmausten keski-
pituuden ja lisaantyvan apuverbien kayton valilla yh-
teys.
Tutkijoiden mukaan tulokset tukevat ajastusta, etta
englanninkielen monimutkainen apuverbijarjestelma
aiheuttaa englanninkielisille lapsille ongelmia.
Saksan kielen apuverbijarjestelma on hyvin paljon
hollannin kaltainen. Kuitenkin Clahsenin tulokset sak-
salaisista lapsista osoittavat apuverbiongelmien olemas-
saolon. Syyna voi olla, etta Clahsenin tutkimuksessa
saksalaisten lasten ilmausten keskipituus
oli
suurempi
kuin hollantilaisten lasten.
Johtopaatos on, etta hollantilaisilla on vahemman
apuverbiongelmia kuin englantilaisilla jarjestelmien
erojen vuoksi. Tutkimus osoittaa kielten valisten vertai-
lujen merkityksen kielellisten hairiiiiden tutkimuksessa.
21