Downloaded from UvA-DARE, the institutional repository of the University of Amsterdam (UvA)

http://hdl.handle.net/11245/2.124415

File ID uvapub:124415
Filename 1: General introduction
Version unknown

SOURCE (OR PART OF THE FOLLOWING SOURCE):

Type PhD thesis
Title Accessing word meaning: Semantic word knowledge and reading
comprehension in Dutch monolingual and bilingual fifth-graders
Author(s) M. Cremer
Faculty FGw: Amsterdam Center for Language and Communication (ACLC)
Year 2013

FULL BIBLIOGRAPHIC DETAILS:

http://hdl.handle.net/11245/1.394952

Copyright

It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or
copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content licence (like
Creative Commons).

UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

(pagedate: 2014-11-25)
Chapter 1

General introduction

1.1 Vocabulary and school success

For children to be successful at school, extensive vocabulary knowledge is needed.
Word knowledge is an important predictor of academic achievement and it plays a
central role in cognitive development, especially in relation to literacy and learning
(Cunningham & Stanovich, 1997; Snow, Burns, & Griffin, 1998). For children the
acquisition of word knowledge is a continuous process that involves learning word
labels, meanings and uses. This happens in parallel with their conceptual
development: as they learn about the world, they learn about words. Although all
children seem to learn words quickly, research shows that there are large differences
between children in lexical knowledge. Disadvantages in word knowledge are found
in particular for bilingual children from an immigrant or minority background whose
home language is different from the language of society (Cito, 2002). Although
under favourable circumstances bilingual children can become highly proficient in
both of their languages, minority children seem to have difficulty coping with the
language of schooling. Differences between children in vocabulary knowledge have
often been linked to differences between children in reading comprehension (cf.
Mancilla-Martinez & Lesaux, 2010; Nakamoto, Lindsey & Manis, 2008; Proctor,
Carlo, August & Snow, 2005), but it is not clear which components of word
knowledge and lexical processing underlie this relationship.
Children’s lexical development involves more than acquiring a large
vocabulary. Besides knowing many individual words, language learners need to
2

know how words are used, which associations go with words and they need to know
about the relationships between words. In addition, for fluent reading learners need
to be able to access this knowledge fluently and recognize and understand words
quickly, as readers construct the meaning of a text on the basis of their
understanding of the words in the text. A piece of text such as Sam will also bring a
hammer. He never forgets his tools will be difficult to process without an
understanding of how the words hammer and tools relate to each other. As school
language and school texts become increasingly complex and abstract, learners’
knowledge of word meaning and of relationships between words becomes ever more
important. Understanding how word knowledge feeds into the reading
comprehension process is important not only from a theoretical perspective. It may
also contribute to the knowledge needed to design better tailored reading programs
to target reading delays.

1.2 Semantic word knowledge

Researchers have traditionally distinguished between two aspects of vocabulary
knowledge: breadth (size) and depth (Anderson & Freebody, 1985). Breadth refers
to how many words are known; depth refers to how well those words are known.
The concept of depth can be defined comprehensively as including a whole range of
dimensions such as spelling, pronunciation, grammatical features, and collocations
(Richards, 1976); it can also more specifically refer to semantic word knowledge in
terms of knowledge of context-independent word meaning and abstract, semantic
relations (Schoonen & Verhallen, 2008; Verhallen, 1994). Although breadth and
depth are closely related (Meara & Wolter, 2004), they are constructs that can be
measured separately (Qian, 1999, 2002; Schoonen & Verhallen, 1998). The
importance of depth or richness of word knowledge may be that, as it increases,
words can be used more flexibly, and their meaning can be readily appreciated and
accessed within multiple contexts (Anderson & Freebody, 1981; Beck, McKeown &
Kucan, 2002).
Several studies have investigated vocabulary depth in terms of knowledge
of semantic relations between words. Knowledge of word meaning develops from
 3

personal, contextual meanings into more abstract, semantic meanings as learners

repeatedly encounter words in language use (Nelson, 2007). As such they build up a
semantic network of word relations or associations. Verhallen (1994; Verhallen &
Schoonen, 1993) found evidence for differences in children’s semantic word
knowledge. She compared 9-11-year-old Turkish-Dutch and monolingual Dutch
children in their assignment of meaning to simple Dutch words. One important
difference between the Turkish-Dutch and the Dutch children was that the Turkish-
Dutch children received less Dutch input at home. On the basis of structured
interviews involving a definition task, she found striking differences in the kind of
knowledge monolingual and bilingual children had of familiar words. The bilingual
children consistently attributed fewer meaning aspects to simple Dutch words and
their knowledge was more context-specific such that they would associate a nose
with context-specific knowledge such as dripping or handkerchief whereas
monolingual children mentioned more semantically related meaning aspects such as
smelling or body part. These differences in associations suggest differences between
children in semantic networks. In a follow-up study, using a standardised paper-and-
pencil word association test, similar differences in semantic word knowledge
emerged and these were found to predict learners’ reading comprehension
(Schoonen & Verhallen, 1998). Qian (1999) investigated vocabulary depth with
young adult learners of English and used a word association format measuring
knowledge of meaning and collocations of presented stimulus words as opposed to
more contextual knowledge. He found that semantic differences between learners
significantly predicted their reading performance beyond the contribution of
vocabulary breadth. These studies show that semantic weakness is associated with
poor reading comprehension (Qian, 1999; Schoonen & Verhallen, 1998).
Since the studies discussed above used non-timed measures of vocabulary
knowledge, they do not reveal much about the role of processing components of
word knowledge. Hence, it remains unclear whether good comprehenders are simply
better at offline tasks such as paper-and-pencil tests or whether they actually have
different semantic networks and process meaning more efficiently, resulting in
better comprehension. Differences in semantic processing between learners have
4

indeed been found (Schreuder & Flores d'Arcais, 1989; Nation & Snowling, 1999,
2004) and such differences are taken to reflect differences in underlying semantic
representations. For example, Nation and Snowling (2004) found that knowledge of
related words in terms of semantic fluency and synonym judgement contributed
unique variance to individual differences in reading comprehension. Considering
such processing differences, a possible explanation for the relationship between
semantic word knowledge and reading comprehension may lie in how word
meaning is processed. It may be that when reading words, less-proficient
comprehenders activate less well-developed concepts and other kinds of related
words than proficient comprehenders, due to differences in their semantic word
knowledge.

1.2.1 Access to semantic word knowledge

Processing aspects of language use have recently become more important in
language learning studies (cf. Segalowitz, 2010). In addition to the roles of breadth
and depth in reading comprehension, several researchers consider fluent processing
as a dimension of word knowledge that may be supportive of reading. Fluency refers
to the ease and speed with which learners can access and use words, as opposed to
simply recognising the words and knowing about how to use them, which is what
more traditional vocabulary tests measure (Daller, Milton & Treffers-Daller, 2007).
Beck and her colleagues (2002) suggested that fluency of access to word
meanings may be a key factor in explaining differences in outcomes of vocabulary
training studies. Vocabulary training that effects the fluency with which word
meanings are accessed has an impact on reading comprehension, but vocabulary
instruction that does not lead to sufficient fluency of access often does not
generalize to reading comprehension (cf. Jenkins, Pany & Schreck, 1978).
What then does lexical fluency entail? There are many basic questions
regarding fluency that are largely unanswered. The terms fluency and automaticity
are used interchangeably in the literature. Based on work in reading theory and
cognitive sciences (Logan, 1988; Perfetti, 1985; Stanovich, 1990), Wolf, Miller and
Donnelli (2000) define automaticity as “a continuum in which processes are
 5

considered automatic when they are fast, obligatory and autonomous and require
only limited use of cognitive resources” (p. 377). They use the term automaticity in
relation to underlying component processes and fluency to refer to fast word
identification and comprehension outcomes. In line with this we could say that
fluent reading comprehension is preceded by fluent word identification, which in
turn may result from automatic activation of word meaning.
It has been suggested that word knowledge and access to that knowledge in
terms of automatic sub lexical processes are interdependent and are a prerequisite
for reading fluency. Initially, in his verbal efficiency theory, Perfetti (1985) posited
that word identification, the rapid retrieval of a word’s phonology and meaning, was
a limiting factor in comprehension. In the more recent lexical quality hypothesis
(Perfetti & Hart, 2001, 2002) the knowledge component is also emphasised: apart
from speed it is important that the reader has the ability to retrieve the meanings that
are needed in a given context. In other words, the quality of the reader’s lexical
representations needs to be sufficient. Perfetti and Hart propose that the quality of
lexical representations drives fast processing and efficient word identification so that
processing resources can be devoted to higher order comprehension (Perfetti & Hart,
2001). The accessibility of word knowledge builds on learners’ underlying semantic
representations and the interconnectedness of lexical-semantic information. Wolf
and colleagues posit that as word meanings become well established through
frequent encounters, they become more quickly accessible (Wolf, Miller, &
Donnelly, 2000). In this thesis we will use the term accessibility to refer to both
speed of access to and activation of word meaning.
A recent study with 203 third-grade readers investigated whether breadth,
depth and fluency of lexical knowledge were distinguishable (Tannenbaum,
Torgesen & Wagner, 2006). The researchers defined fluency as the rate at which
learners access the meaning of a word. For this they administered two tasks: the
Word Use Fluency subtest of the Dynamic Indicators of Basic Early Literacy Skills
(DIBELS; Good & Kaminski, 2002) and an experimenter-developed semantic
category fluency test. The word use fluency test measured the number of target
words correctly used in sentences during a one-minute testing period; in the
6

semantic category fluency test the children named as many items as possible from
eight categories such as farm animals and fruits. Depth was measured using two
definition tests, involving multiple meanings and the naming of attributes.
Tannenbaum and colleagues found that the three dimensions of word knowledge
were not completely distinguishable from each other using the tests employed. A
two-factor model of breadth and depth/fluency provided the best fit to the data. They
found that both breadth and depth accounted for unique variance to reading
comprehension, although the contribution of depth/fluency was small. Over half of
the variance in reading comprehension that was explained by the vocabulary
measures was variance that the two vocabulary factors had in common. The
researchers suggest that depth and fluency are influenced by similar types of
experiences with words and that speed of access to word meaning improves as word
meaning is reinforced and flexibility of use is obtained.

1.2.2 Semantic priming

A more psycholinguistic way of investigating the processing of word meaning is
through semantic priming. Semantic priming refers to the unconscious speeding up
of word recognition when a target word is preceded by a semantically related word.
Priming effects reflect the automatic co-activation of related words. A study by
Nation and Snowling (1999) with 10-11-year-old children compared proficient and
less-proficient readers in the degree of automatic activation of semantically related
words. They measured semantic priming effects in a continuous, auditory lexical
decision task. On average the two groups differed significantly in semantic
processing. Although both groups of comprehenders showed priming for function-
related words (e.g., broom - floor), there were differences in priming for abstractly
related words (category co-ordinates such as airplane - train). For category
coordinates, less-proficient comprehenders only showed priming if the semantic
category pairs were also commonly associated. The researchers conclude that in the
absence of such explicit co-occurrence of words in language use less-proficient
comprehenders are less sensitive to abstract semantic relations. Such findings
suggest that children’s reading comprehension problems may be associated with less
 7

effective semantic processing, or reduced accessibility of semantic knowledge. What

remains unclear here is whether individual differences between children in the
accessibility of word meaning – for example as reflected in semantic priming –
make a unique contribution to explaining variance in their reading comprehension
scores.

1.3 Individual differences between learners

Much research that tries to find out how language works focuses on the fine-grained
cognitive processes underlying language use in an average language user, e.g., how
do we store words in memory, how do we process acoustic signals. At the same time
there is research, partly grounded in educational practice, that focuses on individual
differences between language users, e.g., why are some learners more successful
than others, what type of instruction suits which learners (for an early example of
this tradition, see Cronbach 1957). This individual differences research
systematically compares groups of language users that differ in age, proficiency,
language background, to name but a few variables. As Roberts and Meyer (2012)
point out, the two ‘paradigms’ cannot do without one another. Studying individual
differences is important because any psychological theory of language should be
able to predict and explain differences between language users. Relating differences
in variables such as age or intelligence to differences in learning outcomes is a good
way of studying how several variables together affect a target behavior. In this vein,
Andringa, Olsthoorn, van Beuningen, Schoonen and Hulstijn (2012) compared
variation in native and non-native listening comprehension. Their results showed
that for both groups linguistic knowledge differences explained variation in listening
comprehension, but for the native speakers, processing speed also contributed
substantially, whereas for the non-native speakers there was a significant
contribution of reasoning ability (IQ). Such comparisons are important for
pinpointing whether the contributions of general cognitive components to using
language depend on the speakers’ linguistic proficiency. This thesis focuses on
understanding the relation between subcomponents of lexical knowledge and
reading proficiency. To this end it relates individual differences at the level of
8

lexical knowledge and processing to individual differences at the higher order level
of reading comprehension. In that, this thesis combines an individual-differences
approach and a process-oriented approach.

1.3.1 Bilingual minority children

Bilingual children from a minority background form a population of special interest
for an investigation of the relationship between word knowledge and reading.
Delays in word knowledge and reading comprehension have consistently been
reported for this population. Bilingual minority children often also come from lower
socio-economic backgrounds and it is often hard to distinguish the two factors, let
alone control for these in experimental research.
Differences between children have been reported for vocabulary breadth as
well as depth. First-grade, bilingual Hispanic children in the US have been reported
to have poorer receptive knowledge of L2 words than their (English) first language
(L1) speaking age mates (Umbel, Pearson, Fernandez, & Oller, 1992). In the
Netherlands, children from non-Western immigrant communities on average lag
behind their Dutch peers at school in language subjects as well as mathematics, as
measured with national curriculum tests (Central Bureau voor de Statistiek, 2008).
Regarding language skills, minority children lag behind not only in vocabulary
breadth, but also in knowledge of word meaning. National surveys in grades 1 to 3
show that children with a home language other than Dutch show considerable delays
for reading and listening comprehension as well as knowledge of word meaning and
meaning relations, but not in spelling and sentence building (Cito, 2002). Cross-
sectional data collected from fourth-grade Spanish-speaking and English-only
children from four schools in the US corroborate that bilingual children not only
know fewer English vocabulary words but also that their knowledge of word
meaning is poorer in comparison to their monolingual age mates (August, Carlo,
Lively, Lippman, McLaughlin & Snow, 1999). Importantly, these differences in in-
depth knowledge of word meaning do not always show on the surface.
Parallel to delays in word knowledge significant delays in reading
performance of bilingual children have been signalled even though these children
 9

have experienced the same education as monolingual children (August, Carlo,

Dressler & Snow, 2005). Smits and Aarnoutse (1997) found that, throughout
primary education, bilingual children do not fall behind in decoding skills such as
spelling and word reading, but show poorer performance in tasks such as reading
comprehension and vocabulary.
Research consistently shows that delays in reading comprehension and
vocabulary are difficult to overcome (Biemiller, 2005). A Dutch study (Aarnoutse,
Van Leeuwe, Voeten, & Oud, 2001) reported a two-year delay between proficient
and less-proficient reading comprehenders that persisted throughout primary school.
A Canadian study (Farnia and Geva, 2007) showed that after years of schooling
ethnic minority children do not catch up with their monolingual peers. It is often
difficult for teachers to target these delays since it remains unclear which aspects of
vocabulary knowledge are problematic and how these contribute to reading
comprehension. A Dutch study by Verhallen, Schoonen and Appel (2001) showed
that so-called deep word knowledge is trainable. Eight-to-ten-year old children who
were relatively weak at knowledge of semantic word relations followed a short
training programme (12 hours) focused on three word domains (animals,
occupations, vehicles) and significantly improved their word knowledge in
comparison to a control group. The improvements did not however generalise to
other domains and to reading comprehension.
The reported delays of bilingual minority children for knowledge of word
meaning seem to run parallel to their poorer reading comprehension skills. This
raises the question to what extent individual differences in lexical (semantic)
knowledge and processing can explain individual differences in reading
comprehension. To understand the relationship between these differences, and to
gain an insight into the comprehension problems of these learners, this thesis
investigates to what extent word associations, semantic word knowledge and speed
of access to that knowledge are different in monolingual and bilingual children, in
the Dutch context. Researchers need to understand the precise nature of the
problems of children who struggle with reading comprehension in order to design
10

effective educational interventions and provide literacy education that fits the needs
of both monolingual and bilingual children.

1.4 The organisation of this thesis

The goals of this thesis are to provide an insight into the extent to which underlying
processes help explain the relationship between semantic word knowledge and
reading comprehension, as well as to pinpoint the lexical-semantic differences
between monolingual and bilingual minority children since for the latter a notorious
delay in reading performance is reported. Since research suggests a role for semantic
word knowledge and the accessibility of semantic word knowledge in reading
comprehension, these are components that deserve further investigation. If fast
access to word meaning supports comprehension, differences in accessibility may be
expected to help explain differences in reading performance. We distinguish two
important aspects of semantic word knowledge: availability, the knowledge itself,
and accessibility, the speed with which that knowledge is activated. This
information may help better explain the differences between proficient and less-
proficient comprehenders and between monolingual and bilingual children. We
extend the fluency dimension to unconscious activation of semantic word
knowledge. The research on lexical-semantic processing in an educational context is
still scarce. This thesis contributes to filling that gap.
Chapter 2 provides an overview of the literature relevant to the relationship
between word knowledge and reading and presents the research questions of this
thesis in more detail. Chapter 3 reports on a study investigating qualitative
differences between monolingual and bilingual children and adults in word
knowledge as reflected by their word associations to target words. Chapter 4
presents a study examining learners’ speed of access to semantic word knowledge as
assessed with a speeded categorization task. The contributions of semantic word
knowledge and speed of access to reading comprehension are investigated. Chapter
5 presents a study addressing speed of access by using measures of lexical decision
and semantic classification, and investigates differences between children in
semantic priming. The contributions of speed of access and semantic priming to
 11

reading comprehension are investigated. Together, the empirical studies in Chapters

3 to 5 assess the relation between components of word knowledge and processing
and reading comprehension. Finally, Chapter 6 summarises and discusses the
findings and presents the overall conclusions of this thesis.