Distribution of the Relatives of Children with SLI and their MLU-Matched Peers
Brian Weiler  Magdalene Jacobs  Lisa Wisman Weil  C. Melanie Schuele
Vanderbilt University School of Medicine  Purdue University
INTRODUCTION
RESULTS
PURPOSE
• Understanding the complex syntax profile of children with SLI is
critical to designing appropriate interventions, especially those
related to classroom dialogue.
This study was conducted to determine whether the proportional
distribution of relative types produced by children with SLI
differs from that produced by children with typical language.
• Children with specific language impairment (SLI) have weaker
production of complex syntax than age-matched peers (e.g.,
Barako Arndt & Schuele, 2012; Owen & Leonard, 2006;
Eisenberg, 2004).
Given that children with SLI have difficulties with syntactically complex
productions, the current study specifically considers whether children
with SLI, when using relatives, show greater proportional use of
propositionally less complex structures than their MLU-matched peers.
• Children with SLI struggle with a grammatical requirement of
relative clauses.
• They omit obligatory relative markers to a much greater
extent and for a protracted period of time compared to MLUmatched peers.
• In an elicited language task, children with SLI between 5 7 years of age often omitted obligatory relative markers
(e.g., that) in subject relative clauses (e.g., I like the dog
___ is brown.); whereas, a group of typical language
learners between 3 - 5 years of age never omitted
obligatory relative markers (Schuele & Tolbert, 2001).
• Less understood is the extent to which children with SLI might
also differ from MLU-matched peers in the versatility of their
production of different relative clause types.
• Relative types can be categorized according to the syntactic
role of either the head noun or the relativized noun phrase
(Diessel and Tomasello, 2000).
• The current study focuses on five different types of head
nouns defined in the below table, where cat is the head noun
of the relative clause.
FINDING 1
The proportional distribution of total relative types used by
children with SLI and MLU-matched peers was comparable.
PARTICIPANTS
SLI
M
SD
Range
Age
SPELT-2
CMMS
(Z-score)
(SS)
Finding 2
• The proportion of PN-relatives produced by the MLU group
decreased with age in a manner consistent with that of the four
children with typical language reported on by Diessel and
Tomasello (2000).The development of relatives in the
spontaneous language of children with typical language appears
to follow a pattern of relative clause expansion beyond an early
reliance on PN-relatives.
Number of
Utterances
6;6
10 mos.
5;2 – 7;10
-3.5
2.11
-7.70 to -1.25
99.48
7.15
87-116
200
68.29
97 – 364
M
4;7
-0.01
108.14
207.4
SD
7 mos.
0.52
8.26
61.45
Range
3;2-5;6
-1.00 to 1.00
87 - 124
96 – 364
Percent of Total Utterances Containing Relatives
SLI: 1.7% (73/4200)
MLU: 1.4% (68/4770)
MLU
METHODS
• A logical next step in this line of research would involve
investigation of the spontaneous production of relative types by
children with typical language who are age-matched to the
children with SLI. It is important to determine whether, at the age
of moving into academic language tasks, the range and
versatility with which children with SLI modify noun phrases is
different from their classmates.
FINDING 2
• Within the spontaneous language of typical language learners,
Diessel and Tomasello (2000) reported a 71% to 37% decrease
in the proportion of PN-relatives to all other relative types
between the ages of 3 to 5.
• They speculate that the early proportional frequency of PNrelatives may be related to three factors:
Whereas the proportional frequency of PN-relatives decreased
with age for MLU-matches, a finding consistent with Diessel
and Tomasello (2000), it remained unchanged for children with
SLI.
Children also completed an elicited subject relative clause production
task (adapted from Hamburger and Crain, 1982; Schuele, 1995). The
findings from this task reflect those from Schuele and Tolbert (2001),
namely that children with SLI omit obligatory relative markers more
frequently than younger MLU-matched peers with typical language.
• Children with SLI, on the other hand, did not show an agerelated decline in PN productions. Whether this finding is related
to a prolonged dependence on propositionally simple relative
constructions among children with SLI or is simply related to the
specific pragmatic demands of the conversational context is a
question worthy of future investigation.
FUTURE DIRECTIONS
Relative clause types were classified by consensus between all four
authors using a coding scheme adapted from Diessel (2004) and
Diessel and Tomasello (2000).
• Diessel and Tomasello (2005) posit a bottom-up sequence of
relative clause acquisition which begins with single proposition
constructions where the predicate nominal (PN) of a
presentational copular clause is modified by the relative clause
(e.g., This is the cat that chases mice). Children’s repertoire of
relative productions later expands to more complex
constructions containing two propositions (e.g., The cat that
chases mice eats cheese).
Finding 1
• Although children with SLI struggle with the grammatical
requirements of relative clauses – omitting obligatory relative
markers in subject relatives to a greater extent than MLUmatched peers – they show similar range and versatility in
production across relative types, suggesting relative clause
structure on a broad level may not be problematic.
• Complexity of propositional structure in relatives did not readily
distinguish the spontaneous productions of children with SLI
from MLU-matched peers (PN: χ2 (1) = 2.40, p > .10).
Examiner-child conversational language samples were collected
(Hadley, 1998) and analyzed to study the complex syntax
development of children with SLI.
• 21 children with SLI
• 23 typically developing children matched on MLU (+/-.2)
Group
DISCUSSION
• Longitudinal case studies would permit examination of complex
syntax trajectories over time within subjects rather than between
subjects.
• Future studies should be designed with a methodology that
considers the influence of adult input, propositional complexity,
and pragmatic communicative function on the production of
relative types.
ACKNOWLEDGEMENTS
Data collection was supported by the ASHFoundation and the Schubert Child Development
Center at Case Western Reserve University (PI: Schuele). Completion of this study and
preparation of this poster were supported by a Preparation of Leadership Personnel grant
(H325D080075; PI: Schuele), U.S. Department of Education.
1. PNs are frequently occurring in adult input provided to
children.
2. The pragmatic function of PNs may correspond to early
communicative needs.
3. PNs are less propositionally complex than other relative types
that express relationships between two situations.
LIST OF CITED REFERENCES AVAILABLE ON REQUEST
http://www.mc.vanderbilt.edu/languagelab
Poster Presented at the Symposium on Research in Child Language Disorders
Madison, WI
2013