Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

Reading and Writing: An Interdisciplinary Journal 12: 169–190, 2000.

169
© 2000 Kluwer Academic Publishers. Printed in the Netherlands.

Awareness of the structure and meaning of morphologically


complex words: Impact on reading

JOANNE F. CARLISLE
Department of Communication Sciences and Disorders, Northwestern University, Evanston,
Illinois, USA

Abstract. Many new words middle school children encounter in books they read are relatively
transparent derived forms whose meanings might be figured out through analysis of the word
parts. Of importance is whether students can not only read and recognize the structure of
morphologically complex words but also determine their meanings. This issue was addressed
by investigating the relationship of third and fifth graders’ awareness of the structure and
meanings of derived words and the relationship of these forms of morphological awareness
to word reading and reading comprehension. The results showed that awareness of structure
was significantly related to the ability to define morphologically complex words; some aspects
were also significantly related to the reading of derived words. The three morphology tasks
accounted for significant variance in reading comprehension at both grade levels, but the
contribution was stronger for the fifth than the third grade. It may be educationally note-
worthy that morphological analysis contributed significantly to reading comprehension for
the third graders because they are presumably just beginning to learn to read and understand
morphologically complex words.

Keywords: Linguistic awareness, Morphology, Reading comprehension, Structural analysis,


Vocabulary growth, Wordreading

Introduction

While there is evidence of a relationship between morphological awareness


and reading in the early school years (Carlisle 1995; Carlisle & Nomanbhoy
1993; Champion 1997; Fowler & Liberman 1995), we have only begun the
kinds of investigations needed to understand the nature and extent of this rela-
tionship. The primary focus has been on analysis of children’s knowledge of
morphological relations (e.g., awareness of the composition of inflected and
derived forms). However, since morphemes are units of meaning, a central
issue has been how awareness of word structure is related to understanding
of the meaning of morphologically complex words and how these forms of
morphological awareness are related to reading comprehension.
The present study focuses on three tasks that may require morphological
awareness with the goal of determining how they are related to one another
170 JOANNE F. CARLISLE

and to reading comprehension. The expectation is that morphological analysis


and reasoning (particularly of derived forms) contribute to understanding
written texts if students are aware of the morphological components of words,
as well as their meaning and grammatical roles.

Awareness of morphological structure and meaning

Developing morphological awareness may become very important for readers


in the middle school years. Nagy and Anderson (1984) estimate that 60% of
the new words middle school students encounter are derived forms that have
relatively transparent structures, and the morphemes are familiar enough so
that the reader can make a reasonable guess about the meaning. Similarly,
Anglin (1993) has shown the importance of derived words to vocabulary
growth; in a study designed to estimate the extent of children’s know-
ledge of psychologically basic words and various types of morphologically
complex words, he found that first graders knew some derived forms, but
fewer derived forms than roots words and inflected forms. At the third-grade
level the number of derived forms known by the children was significantly
greater than the number of root words or inflected forms, and this difference
was much more pronounced for fifth graders. Anglin found that morpho-
logical problem solving contributed to students’ performance; for example,
they sometimes mentioned the morphemic constituents and the meanings of
the base morpheme and suffix as a way to figure out the meaning of the
complex word. Such a process might aid students as they encounter unfa-
miliar morphologically complex words in reading (Nagy & Anderson 1984)
and so may contribute to comprehension of texts.
Research on the organization of the mental lexicon and on aspects of
lexical access has suggested that morphological processing works toward
the goal of computing meaning from the constituent elements (Schreuder &
Baayen 1995). Structural analysis alone may be misleading because words
can sound (or look) alike but not be morphological relatives (e.g., bear and
beard). The importance of similarity in meaning is underscored by explora-
tory research on the psychological basis for determining morphological
relations carried out by Derwing (1976). He found that semantic similarity
was more important than phonetic similarity but that phonetic similarity was
more likely to influence judgments of relatedness by elementary children than
high school students or adults.
Morphological awareness, as it contributes to reading, must have as its
basis the ability to parse words and analyze constituent morphemes for
the purpose of constructing meaning. In line with this expectation, some
researchers have reported developmental increases in awareness of morpho-
logical structure and its link to word meanings. Freyd and Baron (1982) found
AWARENESS OF STRUCTURE AND MEANING 171

that able fifth graders were superior to eighth graders at learning the meaning
of derived forms and inferred that this was because the fifth graders were
better able to figure out word meaning from analysis of the base words and
suffixes. Both groups tended to base their definitions on the base words,
often ignoring or misrepresenting the meaning of the suffixes. Somewhat
similarly, Wysocki and Jenkins (1987) investigated fourth, sixth, and eighth
graders’ ability to use morphological and contextual information to figure
out the meaning of unfamiliar words. The students’ success in figuring out
the meaning of such words depended on prior experience with related words
and the strength of the surrounding sentence context. Fourth graders were less
able than the older students to use both types of information to figure out word
meanings. Like Freyd and Baron, these researchers found that the students
were likely to define words in terms of their base forms, without regard for the
meanings of the suffixes. Additional evidence of developmental trends comes
from Tyler and Nagy (1989), who found that knowledge of the syntactic
properties and meanings of suffixes was limited among fourth graders but
less so by eighth grade. Similarly, Windsor (1994) found that comprehension
and production of derivational suffixes was significantly better for older than
for younger students.
Developmental changes in students’ grasp of the relation of the struc-
ture and meaning of derived forms may be affected by the transparency of
the structure and the productivity of the suffix (e.g., Carlisle 1988; Cham-
pion 1997; Tyler & Nagy 1989). Transparency and productivity have been
found to be related to successful interpretation and formation of derivatives
in meaningful contexts. Carlisle (1988) found that, when errors were made in
producing derived forms, younger students often used common, productive
suffixes that did not require phonological changes (e.g., given produce and a
sentence context, the student said producement instead of production).
In order to gather more direct evidence of the relation of awareness of
structure and meanings of morphologically complex words, the present study
included tasks of both structural analysis (decomposition and derivation of
forms) and definition. The assumption was that the derivation task would be
more closely related to students’ ability to define morphologically complex
words than the decomposition task because producing derived forms, like
defining derived forms, requires knowledge of the grammatical roles and
meanings of suffixes, not just relational knowledge. Additionally, because
defining morphologically complex words may involve analysis of the mean-
ings and grammatical roles of the morphemic constituents, this task may tap
processes similar to those students might use when they encounter morpho-
logically complex words in reading. Snow (1990) and other (e.g., Litowitz
1976) have suggested that defining words requires metalinguistic capa-
172 JOANNE F. CARLISLE

bility; students must treat words as objects of thought, integrating linguistic


properties and contexts that they call up from memory.

Morphological awareness and reading

Along with investigating links between awareness of structure and meaning,


the present study was designed to determine whether morphological aware-
ness is related to reading achievement. This included word reading, assessed
by the students’ ability to read derived forms, and reading comprehension.
One question was whether structural analysis was related to the reading of
derived words. The other was whether morphological awareness (in terms
of structure and meaning) and the reading of derived words contribute to
reading comprehension. In terms of adult word reading, researchers have
shown that morphological processing plays a role, although the organiza-
tion of the lexicon with regard to storage and processing of morphemes
continues to be debated (e.g., Chialant & Caramazza 1995; Cole, Beauvillain
& Segui 1989). Recently, Napps (1989) found that morphemic relations are
represented explicitly in the lexicon. Cole et al. (1989) showed that suffixed
words are accessed via the word root for French adults; the members of a
family are then examined in a frequency-ordered search. Whether the same
processes occur in children’s reading of morphologically complex words is
an open question. Tyler and Nagy suggest that there may be a trade-off
between morphological processing and speed of word recognition. In their
words: ‘It is possible that morphological complexity speeds the process of
access to the point at which the set of words which contain base morphemes
is accessed, but that further operations involved in distinguishing related
forms slow subsequent processing. Thus, morphological complexity appears
to facilitate some aspects of lexical access and complicate others, without
necessarily resulting in a net quantitative difference in processing difficulty’
(1990: 31).
One way to investigate the effects of morphological structure on word
reading is to compare performance on sets of words that differ on character-
istics that might affect awareness of structure. One such comparison involves
frequency; if high and low frequency transparent derived forms (both sets
having high frequency base forms) are read with comparable accuracy, it
is possible that readers are basing their reading of the words on awareness
of the base forms (all with high frequency). Because word frequency is
generally very strongly related to accuracy of word reading, such a finding
might be evidence of morphological decomposition and use of the base form
to pronounce the word. Another comparison involves linguistic complexity;
transparent derived words would presumably be easier to decompose than
derived words with phonological and orthographic shifts. If students give
AWARENESS OF STRUCTURE AND MEANING 173

less accurate responses to shift words that in other respects do not differ
from transparent word (e.g., frequency, length), a likely reason might be the
opaqueness of the morphological structure. This effect has been shown on
shift words on oral tasks in numerous studies (e.g., Carlisle 1988; Champion
1997; Fowler & Liberman 1995; Jones 1991; Leong 1989; Tyler & Nagy
1989).
Because developmental trends suggest relatively limited awareness of
the structure and meaning of derived forms until after the fourth grade
(e.g., Carlisle 1995; Carlisle & Nomanbhoy 1993), performance of younger
and older middle-school students might differ significantly on a task of
reading derived forms. The younger students are likely to be better at decom-
posing than producing complex forms and to be still learning basic strategies
for ‘sounding out’ polysyllabic words. They might not use morphological
analysis in their efforts to read derived forms. They would be expected
to read high frequency words more accurately than low frequency words,
regardless of the familiarity of the base form; similarly, transparency of the
morphological structure might not affect their reading of derived forms. Older
students, on the other hand, might have sufficient experience with the ortho-
graphic representation of morphologically complex words so that they notice
morphemic constituents and use them in reading the words. Thus, they might
read transparent words more accurately than those that undergo phonolo-
gical and orthographic changes; they might also read unfamiliar transparent
words as accurately as familiar transparent words because of their growing
experience with systematic relations in word formation (e.g., Chomsky 1970;
Templeton & Scarborough-Franks 1985).
Children’s skill at reading morphologically complex words has been
directly assessed in only a few studies. Brittain (1970) found that know-
ledge of inflected forms on an oral task was related to reading achievement.
Elbro and Arnbak (1996) investigated reading of compounds (compared to
other words) by Danish dyslexic adolescents and reading-age matched normal
readers. They found that the dyslexics were more sensitive to or dependent
on morphological structure than the normal readers; thus, in Danish more
skilled readers might not depend on morphological segmentation, having
developed automatic recognition of words as entities. Fowler and Liberman
(1995) looked at the contribution of morphological awareness on an oral task
to word reading, but the word reading test was a standardized measure of
general word reading. Nonetheless, they found that the ability to recover the
base morpheme from a derived form that was phonologically complex (e.g.
finding nature in natural) was predictive of word reading, after the effects of
age and receptive vocabulary were accounted for.
174 JOANNE F. CARLISLE

Several other studies have shown that good and poor readers differ in
their awareness of morphological structure (Champion 1997; Leong 1989)
but have not provided direct evidence that awareness of morphological
structure matters in word reading or that the ability to read morphologic-
ally complex words is related to reading comprehension. Yet others have
looked at comprehension, but without a measure of word reading. Tyler and
Nagy (1990) studied students’ interpretations of derived words in sentence
contexts. They found that the high school students made fewer lexical errors
on suffixed than non-suffixed words, suggesting some effect for morpholo-
gical processing of words. In keeping with the results of their earlier study
(1989), they found that the students made more syntactic errors on suffixed
than non-suffixed words, suggesting incomplete processing of the affix and its
meaning. Mahoney (1994) found that morphological sensitivity was related
to reading comprehension for ninth graders and adults.
The relationship of skill at reading derived forms and reading compre-
hension is a central component of the present study; middle school students
(third and fifth graders) were chosen as participants because they presumably
have acquired basic reading skills. Third and fifth graders differ in terms their
exposure to complex words in written texts. The expectation was that third
graders would be able to decompose transparent derived forms and show
awareness of the meaning of base morphemes. For them, the contribution
of morphological analysis and reading of derived words to comprehension
might be quite modest. In contrast, for the fifth graders performance on the
morphology tasks was expected to contribute significantly to reading compre-
hension. Exposure to derived forms in reading should have heightened their
sensitivity to the orthographic redundancy of morphologically related words
and provided opportunities to learn strategies for analyzing words (Templeton
& Scarborough-Franks 1985).
The research questions were (1) What is the relationship between chil-
dren’s knowledge of the meanings of morphologically complex words and
their awareness of word structure and use? Does awareness of structure
contribute significantly to their definitions of morphologically complex
words? (2) Is there a significant relationship between awareness of morpho-
logical structure and reading of morphologically complex words? and (3)
To what extent do awareness of morphological structure, the ability to read
morphologically complex words, and knowledge of the meanings of morpho-
logically complex words contribute (together and independently) to reading
comprehension?
AWARENESS OF STRUCTURE AND MEANING 175

Method

Subjects. The subjects were 34 third graders (18 boys and 16 girls) and 26
fifth graders (10 boys and 16 girls) who attended a private day school in a
middle- to upper middle-class suburban community in the midwest. Because
of missing data, one fifth grader is not included in the data analysis, leaving
25 fifth graders. During this school year, the school provided special help
in reading and language arts for 8 of the third graders and 6 of the fifth
graders. None of these students had Vocabulary or Comprehension scores
below the 30th percentile on the achievement battery administered by the
school, the Comprehensive Testing Program (Educational Records Bureau).
Mean percentiles on this test battery for the third graders were Vocabulary
76.8 (15.8 SD) and Comprehension 74.9 (22.4 SD); for the fifth graders,
Vocabulary 80.8 (22.4 SD) and Comprehension 79.2 (13.3 SD).

Word Reading Test (WRT). The purpose of this test was to assess students’
ability to read morphologically complex words. It included three groups of
words, as shown in the Appendix. Set A was made up of 25 words that have
high surface frequency; that is, the Standard Frequency Index was at or above
40.1 Set A contained 9 words in which the base form was fully represented
in the derived form (orthographically and phonologically), referred to here
as Transparent words (e.g., movement, powerful) and 16 words which were
different from their base forms in terms of orthographic and/or phonolo-
gical characteristics (referred to here as Shift words) (e.g., natural, easily).
The suffixes were common and productive for young school-age children,
including nouns (-ment, -ness), adjectives (-y, -ful), and adverbs (-ly). The
frequency (using the Standard Frequency Index) of base and surface forms
was similar: the mean for base words was 57.6 (6.2 SD) and for derived
forms was 56.8 (4.0 SD). Set B was made up of 20 words that have high
frequency base – a mean of 56.8 (4.0 SD) – but low surface frequency – a
mean of 33.3 (1.5 SD). All of the words on Set B had transparent relations
(e.g., puzzlement). The words in Sets A and B were comparable in the base
word frequency and word length; the mean number of letters for Set A was
7.9 (1.7 SD) and for Set B was 8.1 (1.2 SD).

Test of Morphological Structure (TMS). This experimental measure was


designed to assess students’ awareness of the relations of base and derived
forms. One part required the decomposition of derived words in order to finish
sentences (henceforth called Decomposition), and a second part required the
production of a derived word in order to finish a sentence (henceforth called
Derivation). The test items are given in the Appendix. The base and derived
forms were equivalent in word frequency on the two tasks: for Decomposi-
176 JOANNE F. CARLISLE

tion, the mean SFI for the base forms was 56.7 (5.0 SD) and for the derived
forms was 50.2 (5.9 SD), while for Derivation, the mean SFI for the base
forms was 55.8 (6.2 SD) and for the derived forms was 51.0 (5.5 SD).
The two tasks contained equal numbers of word relations that are Trans-
parent (i.e., the sound of the base form is intact in the derived form, as
in reason and reasonable) and Shift words (i.e., the phonological repres-
entation shifts from base to derived form, as in produce and production).
Suffixes judged to be familiar to third and fifth graders were used; these
included -th (e.g., growth), -ance/ence (e.g., performance), -er (e.g., teacher),
-ity (e.g., equality), -tion/sion (e.g., description), -ous (e.g., famous), -able
(e.g., variable). Each suffix was equally represented on the two tasks.

The Test of Absolute Vocabulary Knowledge (TAVK) (Anglin 1993). The


TAVK involves a word interview in which the child is asked to give the
meaning of a word, to use it in a sentence, and (when needed) to pick the
meaning from a multiple-choice set. The words were ranked by Anglin in
order of difficulty. Morphologically complex words in his sample of words
from an unabridged dictionary included inflected forms (e.g., reports, sourer),
derived forms (e.g., knotless), compounds (e.g., milk cow), and idioms. The
idioms were not included in the analysis for this study because, like psycho-
logically basic words, the meaning of an idiom can not be derived from
its constituent morphemes. The reader is directed to Anglin (1993) for a
complete explanation of the development and administration of this test and
for the word list. For this study, the criterion for discontinuing testing was
different from Anglin’s; the students were tested until they said they did not
know the meaning of seven sequential words, five of which they also failed
on the multiple-choice component. The method for assigning credit for a
definition and sentence also differed from Anglin’s. Students received credit
if they explained the meaning of the base word and used it appropriately in
a sentence. The students often were unable to define suffixes because of the
high degree of linguistic abstraction involved; the correct use of a word in a
sentence showed that the child knew the grammatical role and appropriate
usage of word. For purposes of the present study the variable of interest
was the number of morphologically complex words for which the student
earned combined credit for definition and sentence, hereafter referred to as
Definition. The following examples show relevant scoring criteria.
Examiner: What does enjoyable mean? CD: Enjoyable means fun. Like,
I enjoy reading. I like to read. Examiner: Can you use enjoyable in a
sentence? CD: Scott thought reading was enjoyable.
Examiner: What does the word sourer mean? WH: I know what sour
means. Say I have a sour lemon, but then I tried something much more
sour. That means it’s not sweeter and much more sour. Examiner: What
AWARENESS OF STRUCTURE AND MEANING 177

does sour mean? WH: Like a lemon is really sour. It doesn’t taste very
good. It makes your mouth go like this [puckered her mouth]. Examiner:
Can you use sourer in a sentence? WH: The lemon was sourer than the
lime.
Note that CD received Definition credit because he explained the meaning of
enjoy, and used enjoyable in a sentence, while WH received Definition credit
because she showed the meaning of the comparative form of the adjective
as she explained the word (i.e., ‘much more sour’) and also used the form
correctly in a sentence.

Comprehensive Testing Program (CTP), Vocabulary and Reading Compre-


hension subtests. Two subtests from the Comprehension Testing Program
III (Educational Records Bureau) were used. The Vocabulary subtest entails
reading a word and then selecting the best meaning from a choice of options.
The Reading Comprehension subtest entails reading short passages and
selecting the best answers to comprehension questions. The appropriate form
for third and fifth graders was administered to each group.

Procedures. During the winter term, each student was given the WRT, the
TMS, and the TAVK individually in a quiet setting. Administration of the
WRT and the TMS took about 20 minutes; administration of the TAVK took
place in 20-minute sessions on different days until the student reached the
discontinuation criteria. The CTP was administered to each class by school
personnel in the spring of the school year.
Performance on the WRT was tape-recorded and scored by the researcher
or a research assistant. The student received two points for reading a word
correctly within two seconds, but one point if there was a delay of more
than two seconds before reading the word, a false start, a self-correction, or a
repetition of the word. Five protocols from each grade were scored a second
time from the tape by the researcher. Agreement of the two scorings exceeded
95% at each grade level.
Performance on the TAVK was scored by adding the total points for
morphologically complex words that received Definition credit. The students
responses were scored during administration but also tape-recorded. One
fourth of the transcriptions of students’ protocols were scored by both the
researcher and a research assistant. Interrater reliability was 0.93 for the fifth
grade and 0.96 for the third grade.
178 JOANNE F. CARLISLE

Results

Grade-level performance on experimental measures

Table 1 shows the performance of the third and fifth graders on the experi-
mental measures. The fifth graders performed significantly better than the
third graders on the Definition task, t(58) = 4.36, the WRT, t(58) = 4.32,
and the TMS, t(58) = 5.64, all p < 0.001. The group means show that the
fifth graders approached the ceiling on several parts of the TMS and WRT,
specifically the Decomposition task on the TMS and the high frequency
Transparent words (Set A) on the WRT.
For the third grade on the TMS, analysis of variance with two within-
subjects factors (task and level of linguistic complexity) was carried out.
The results showed a significant effect for task; Derivation was weaker than
Decomposition [F(1,33) = 144.37; p < 0.001]. There was also a significant
effect for linguistic complexity, as Transparent words were responded to more
accurately than Shift words [F(1,33) = 167.99; p < 0.001]. The interaction was
also significant [F(1,33) = 226.68; p < 0.001]. Paired contrasts show signi-
ficant effects for linguistic complexity on each task (Transparent > Shift); in
addition, the measures of linguistic complexity differed significantly across
tasks (for Shift, Decomposition > Derivation; for Transparent, Decomposition
> Derivation).
For the fifth grade, because tests of heterogeneity of variance were signi-
ficant due to ceiling effects, performance on the TMS was analyzed using
Friedman Analysis Of Variance for Ranks; results showed that the students
performed differently on the levels of linguistic complexity of the two tasks
[χ 2 (3, N = 25) = 52.76; p < 0.001]. Follow-up tests showed significant
differences in linguistic complexity for each task (Transparent > Shift) and
each type of linguistic complexity across tasks (as was true for the third
grade).
On the WRT, tests of heterogeneity of variance for the three groups of
words were significant. Therefore, comparison of performance at each grade
level on the three word types was carried out using nonparametrics. Friedman
ANOVA for Ranks for the third grade showed that students’ reading of the
three types of words differed significantly [χ 2 (2, N = 34) = 42.61; p < 0.001].
Follow-up analysis showed that the high surface frequency Transparent and
Shift words of Set A differed significantly [χ 2 (1, N = 34) = 22.53; p < 0.001]
and the Transparent words of Sets A and B also differed significantly [χ 2 (1,
N = 34) = 30.12; p < 0.001]. Similarly, the fifth grade students’ reading of
the three types of words differed significantly [χ 2 (2, N = 26) = 33.40; p <
0.001]. Follow-up tests showed that their performance on the high surface
frequency Transparent and Shift words differed significantly [χ 2 (1, N = 26) =
AWARENESS OF STRUCTURE AND MEANING 179
Table 1. Performances on experimental morphology tests by third and fifth graders

TMS WRT DEF


Deriv Decomp Set A Set A Set B
Trans Shift Trans Shift Trans Shift Trans

Grade 3 75.2 38.5 85.1 79.2 92.8 77.2 72.1 20.4


(14.6) (7.4) (11.6) (12.2) (10.2) (19.3) (17.0) (8.6)

Grade 5 88.6 63.1 96.1 91.7 97.1 94.2 85.5 32.3


(9.0) (10.1) (5.1) (10.0) (3.0) (5.6) (9.1) (13.1)
TMS = Test of Morphological Structure (percent correct); WRT = Word Reading Test
(percent correct); DEF (raw score); Deriv = Derivation; Decomp = Decomposition; Trans
= Transparent.

11.84; p < 0.001]; and the high and low frequency Transparent words differed
significantly [χ 2 (1, N = 26) = 22.15; p < 0.001].

Awareness of structure and meaning

The first question concerned the relationship between students’ awareness


of structure and knowledge of the meanings of morphologically complex
words. Pearson correlations (shown in Table 2) of TMS and Definition were
significant: for grade 3, r = 0.46, p < 0.01, and for grade 5, r = 0.64, p <
0.001. Standard regression analyses showed that the two tasks on the TMS
accounted for a significant 26% of the variance in Definition [for the third
grade, F(2,31) = 5.43; p < 0.01]. As is shown in Table 3, only the Derivation
task contributed significantly to the equation. For the fifth grade, the two tasks
on the TMS accounted for a significant 43% of the variance in Definition
[F(2,22) = 8.14; p < 0.01]. Here too only Derivation contributed significantly.

Awareness of structure and reading morphologically complex words

The second question concerned the relationship between awareness of


morphological structure and reading of morphologically complex words.
Table 4 shows the correlations of transparent and shift words on the two
tasks. Only a few of the correlations were significant for either grade level.
At both grade levels performance on Shift words on the Derivation task was
significantly correlated with reading of Shift words on the WRT. For third
graders performance on the Decomposition task was significantly related to
reading high frequency shift words and low frequency transparent words on
the WRT.
180 JOANNE F. CARLISLE

Table 2. Correlations of performances on morphology and reading tests for


third and fifth graders

TMS WRT Definition Voc Rdg. Com

Grade 3
TMS –
WRT 0.36* –
Definition 0.46** 0.18 –
Voc 0.40* 0.54*** 0.42** –
Rdg Com 0.18 0.61*** 0.24 0.54*** –

Grade 5
TMS –
WRT 0.39* –
Definition 0.64*** 0.40* –
Voc 0.61*** 0.57** 0.58** –
Rdg Com 0.69*** 0.39* 0.63*** 0.71*** –
TMS = Test of Morphological Structure; WRT = Test of Word Reading; Voc =
Reading Vocabulary; Rdg Com = Reading Comprehension.
* p < 0.05; ** p < 0.01; *** p < 0.001.

Table 3. Contribution of awareness of morphological structure to the defini-


tions of morphologically complex words

Variable Beta t p-level

Grade 3
Derivation 0.50 2.18 0.037
Decomposition 0.02 0.07 0.945

Grade 5
Derivation 0.61 2.49 0.021
Decomposition 0.05 0.22 0.825

Contribution of morphology tasks to reading achievement

The third question concerned the extent to which morphological awareness


(structure and meaning) and reading derived forms contribute to reading
comprehension at the word level (Vocabulary) and the text level (Comprehen-
sion). Two standard regression analyses were carried out at each grade level.
In each, the independent variables were TMS, WRT, and Definition. The
AWARENESS OF STRUCTURE AND MEANING 181
Table 4. Correlations of aspects of the test of morphological structure
and word reading test for third and fifth grades

TMS WRT
Set A Set A Set B
Trans Shift Trans

Grade 3
Deriv Trans 0.12 0.28 0.13
Deriv Shift 0.25 0.41* 0.23
Decomp Trans 0.26 0.37* 0.36*
Decomp Shift 0.11 0.40* 0.10

Grade 5
Deriv Trans −0.01 0.24 0.14
Deriv Shift 0.30 0.41* 0.30
Decomp Trans −0.04 0.01 −0.08
Decomp Shift −0.03 0.21 0.47*
TMS = Test of Morphological Structure; WRT = Word Reading Test;
Deriv = Derivation task; Decomp = Decomposition task; Trans =
Transparent.
* p < 0.05.

dependent variables were Vocabulary and Reading Comprehension from the


CTP battery. Table 2 shows the correlations of these variables at each grade
level. Performances on all three morphology tasks significantly correlated
with reading achievement for the fifth grade, whereas TMS and Definition
were not significantly related to Reading Comprehension for the third grade.
For the third grade, the results showed that the three morphology tests
accounted for 41% of the variance in Vocabulary [F(3,29) = 6.59; p < 0.01].2
Table 5 shows that the WRT was the only variable that made a significant
contribution. The squared semipartial correlations showed that the unique
contribution of WRT was 16.8%, whereas Definition was 6.8% and TMS
was negligible. For Reading Comprehension, the three morphology tests
accounted for 43% of the variance [F(3,30) = 7.42; p < 0.001]. Table 5 shows
that WRT made the only significant contribution. Squared semipartial correla-
tions showed that WRT made a unique contribution of 34.8%; Definition was
2%, and TMS was negligible.
For the fifth grade, the results showed that the three morphology tests
accounted for 53% of the variance in Vocabulary [F(3,21) = 8.03; p < 0.01].
Table 5 shows that WRT was the only variable that made a significant contri-
bution. Squared semipartial correlations showed that the unique contribution
182 JOANNE F. CARLISLE

Table 5. Contribution of morphology tests to reading vocabulary and


comprehension

BETA t(29) p-level

Grade 3
For Vocabulary:
TMS 0.10 0.59 0.560
Definition 0.30 1.83 0.077
WRT 0.44 2.89 0.007

For Reading Comprehension:


TMS −0.12 −0.72 0.477
Definition 0.16 1.02 0.316
WRT 0.64 4.25 0.000

Grade 5
For Vocabulary:
TMS 0.34 1.75 0.090
Definition 0.23 1.23 0.233
WRT 0.35 2.10 0.048
For Reading Comprehension:
TMS 0.48 2.54 0.019
Definition 0.30 1.60 0.124
WRT 0.09 0.58 0.566

TMS = Test of Morphological Structure: WRT = Word Reading Test.

of WRT was 9.6%, whereas Definition was 3.2% and TMS was 6.8%. In
terms of Reading Comprehension, the three morphology tests accounted
for 55% of the variance [F(3,21) = 8.68; p < 0.001]. Table 5 shows that
TMS made the only significant contribution. Squared semipartial correlations
showed that TMS made a unique contribution of 13.7%; Definition was 5.3%,
and WRT was negligible.

Discussion

The results of the study fill in some missing connections in our understanding
of the value of morphological awareness for reading. Previous research
has shown that students at or above the fourth grade are quite skilled at
decomposing morphologically complex words, particularly when they have
a transparent relation to the base form, and using the base form, at least, to
AWARENESS OF STRUCTURE AND MEANING 183

define derived forms. The results of the present study add to this literature
by providing evidence that for third and fifth graders there is a significant
link between awareness of structure and the ability to define morphologically
complex words. In terms of the impact on reading, previous studies had not
directly assessed the link between relational knowledge and the ability to
read derived forms. The results of the present study show that these aspects
of morphological processing are significantly related. Finally, the study was
designed to determine the extent to which morphological awareness (struc-
tural knowledge and meaning) and the ability to read derived forms contribute
to reading comprehension. At both grade levels, the morphology measures
together contributed significantly to reading comprehension. The relation-
ships were particularly strong for the fifth graders, but it is noteworthy that
they were significant for the third graders, who are presumably still learning
basic strategies for recognizing polysyllabic words in print. While the results
appear to help us understand the importance of morphological awareness and
processing, they must be interpreted cautiously because of the relatively small
groups and because of ceiling effects for the fifth graders.

Morphological structure and meaning

The task used to assess students’ awareness of the morphological structure


of words placed emphasis on their ability not only to find base morphemes
within complex words (e.g., produce within production) but also to produce
complex forms, given the base word and a sentence context. Both third
and fifth graders were quite proficient at decomposing the complex words;
they were significantly better at decomposing complex forms than deriving
them. Both tasks contained derived words that were transparent renditions
of the base word and others that underwent phonological and orthographic
shifts. Here too significant differences were found: across the two tasks,
performance on shift words was significantly less accurate than on trans-
parent words. These results confirm the suggestion by Tyler and Nagy (1989)
that even before fourth grade, students are likely to have sufficient aware-
ness of the structure of derived forms to find the base morphemes when the
phonological representation is intact. While there was a significant effect for
linguistic complexity on each task, performance on the shift words was much
stronger on the Decomposition than on the Derivation task. Even for fifth
graders, producing the correct derived form when phonological representa-
tions change is a difficult task, as other studies have also found (e.g., Carlisle
1988; Champion 1997; Leong 1989; Tyler & Nagy 1989).
Because a central issue in this study was the link between structure know-
ledge and meaning, an important finding was the significant relationship
between structural awareness and Definition performance at both grade levels
184 JOANNE F. CARLISLE

(0.64 for the fifth grade and 0.46 for the third grade). A regression analysis
showed that at both grade levels it was the Derivation task that contributed
most significantly to the students’ definitions of complex forms. This makes
sense, given the nature of the Definition task, since students received credit
only when they could both define and use the word correctly. They needed to
have knowledge of the meaning and grammatical role of the suffix, not just
the meaning of the base word.
Because of the nature of the Definition task, the students’ credit for
Definitions include some instances of what Anglin (1993) called morpho-
logical problem solving, as well as some crystallized knowledge about the
meaning of certain complex words. Students received Definition credit when,
by analysis of the parts, they worked out the meaning of words that they
previously may not have been able to define, such as ‘fenderless’. Certainly
not all derived words lend themselves to this sort of analysis; nonetheless,
morphological analysis was used by both third and fifth graders to generate
possible meanings of some words.

Morphological awareness and reading

Most studies have looked at the relationship between morphological aware-


ness and reading achievement indirectly (e.g., Mahoney 1994; Tyler & Nagy
1990). In the present study an effort was made to determine whether morpho-
logical processing was used to read aloud derived forms. Students’ reading of
high and low frequency transparent derived words (matched for the frequency
of their base forms) was used to see whether fifth graders, in particular, would
be equally accurate on both, because of the familiarity of the base forms.
This result might provide evidence of morphological decomposition prior to
pronunciation of the words. A second comparison involved words for which
the relation of base and derived forms was either transparent or characterized
by phonological and/or orthographic shifts.
In terms of frequency, the result was that high frequency transparent
words were read more accurately than low frequency transparent words at
both grade levels. Thus, the comparable frequency of the base forms did not
offset the differences in surface frequency. However, the ceiling effect for the
fifth graders makes it difficult to interpret the results. Another and perhaps
relevant finding was that performance on the high frequency words was not
significantly related to awareness of structure on the oral test (TMS). While
it is possible that the students did not parse the words into morphemes for
purposes of pronunciation, the very simple measure of automaticity of word
reading used in this study (i.e., two points for an immediate, correct reading,
versus one point for a delayed or corrected reading) may not be sensitive
AWARENESS OF STRUCTURE AND MEANING 185

enough to show the influence of morphological processing. Certainly, a more


fine-grained measure of word reading speed should be used in future studies.
In terms of linguistic complexity, both third and fifth graders performed
significantly less well on shift words than transparent words. Thus, their
reading appears to have been affected by the nature of morphological rela-
tions. An additional finding is that at both grade levels oral production of shift
words on the TMS was significantly related to the reading of shift words. This
would seem to suggest awareness of the complex relations of the shift words
for reading. Another point of interest is that for the third graders, decomposi-
tion of transparent and shift words was also significantly related to the reading
of shift words. Fowler and Liberman (1995) found that it was particularly
the recognition of base forms with complex phonological representations that
was predictive of word reading achievement. While somewhat different, the
results of both studies suggest that structural awareness contributes to the
reading of derived words, although for a more complete understanding of
morphological processing during word reading, measures of word reading
speed as well as accuracy should be included.
As noted earlier, it is possible that reading experience helps students
develop an awareness of morphological relations. While the study reported
here was not designed to investigate this possibility, a post-hoc analysis
was carried out to determine whether grade-level differences in structural
knowledge and the ability to define morphologically complex words would
disappear when the effects of reading ability were statistically controlled. The
results showed that for both tasks the grade level differences were still signi-
ficant when performance on the WRT was used as a covariate [for Definition,
F(1,56) = 7.40; p < 0.01, and for the TMS, F(1,56) = 12.93; p < 0.001]. Thus,
growth of morphological awareness, in terms of structure and meaning, can
not be solely attributed to learning to read derived words.
In terms of reading comprehension, some studies (e.g., Mahoney 1994)
have shown significant correlations between sensitivity to morphological
structure and reading achievement, but without a measure of the ability to
read derived words as a direct link. The results of the present study showed
that awareness of structure and meaning and the ability to read derived words
contributed significantly to comprehension at the word and text levels at both
grade levels. For the third grade the three morphology tasks accounted for
41% of the variance in Vocabulary and 43% of the variance in Compre-
hension. The word reading test made the major contribution in both cases;
its unique contribution to Vocabulary was 16.8% of the variance and to
Comprehension was 34.8%. Definition made a small but significant unique
contribution to Vocabulary (6.8% of the variance). For the fifth grade, the
three morphology tasks accounted for 53% of the variance in Vocabulary
186 JOANNE F. CARLISLE

and 55% of the variance in Comprehension. The unique contributions to the


Vocabulary test were very small for all three variable – the word reading test
accounted for 9.1% of the variance and the TMS accounted for 6.8%. For
comprehension, the TMS accounted for a unique 13.7% of the variance, and
Definition accounted for 5.3%. An important finding, then, is that for both
reading measures, a large portion of the variance accounted for by the three
morphology measures was shared. The morphology tasks tap processes that
are strongly related to one another; together they have a stronger impact on
comprehension than any one would by itself, whether at the word or text level.
The interrelation of different aspects of morphological awareness might
be expected for older students. They have had greater exposure to complex
words in print and more opportunity to learn to use morphological decom-
position and problem solving as an aid to reading than younger students.
However, it is rather surprising that the significant interrelations were also
found for third graders. Previous studies have shown that fourth graders’
morphological awareness may be limited by their tendency to focus primarily
on base morphemes, neglecting the grammatical roles and meanings of
suffixes and the complex phonological and orthographic relations (Carlisle
1988; Tyler & Nagy 1989; Tyler & Nagy 1990; Windsor 1994; Wysocki
& Jenkins 1987). Though morphological awareness was clearly less well
developed for third than fifth graders, it was nonetheless significantly linked
to reading achievement.
Further study is needed to address causal relations. The ability to decom-
pose words into constituent morphemes, evident among third graders, may
precede but may also facilitate the ability to work out the meanings of both
base forms and suffixes. It is likely that exposure to the printed word also
influences the development of morphological awareness. Skilled readers,
because they read more books and more challenging books than less skilled
readers, might use orthographic transparency to enhance their recognition of
the morphological structure of words. The results of the present study cannot
confirm these inferences about reciprocal causal relations but would support
them.

Acknowledgments

I would like to thank the students, teachers, and administrators at the Lake
Forest Country Day School for their assistance in this research project and
Jerry Anglin for his generosity in sharing his research materials and his
advice.
AWARENESS OF STRUCTURE AND MEANING 187

Notes
1. For purposes of matching word sets in terms of frequency in written texts, logarithmic
transformations of the frequency ratings from Carroll, Davies and Richman (1971) were
used; a value of 50 represents a word that appears once in 100,000 words of text. ‘High’
ratings for surface frequency ranged from 50 to 63; ‘low’ ratings (Set B) ranged from 30
to 36.
2. Vocabulary score for one third grader is missing.

Appendix 1: Test of morphological structure

Part 1: Derivation
Practice a. Farm. My uncle is a . [farmer]
b. Help. My sister is always . [helpful]
1. warm. He chose the jacket for its . [warmth]
2. teach. He was a very good . [teacher]
3. permit. Father refused to give . [permission]
4. profit. Selling lemonade in summer is . [profitable]
5. appear. He cared about his . [appearance]
6. express. ‘OK’ is a common . [expression]
7. four. The cyclist came in . [fourth]
8. remark. The speed of the car was . [remarkable]
9. protect. She wore glasses for . [protection]
10. perform. Tonight is the last . [performance]
11. expand. The company planned an . [expansion]
12. revise. This paper is his second . [revision]
13. reason. Her argument was quite . [reasonable]
14. major. He won the vote by a . [majority]
15. deep. The lake was well known for its . [depth]
16. equal. Boys and girls are treated with . [equality]
17. long. They measured the ladder’s . [length]
18. adventure. The trip sounded . [adventurous]
19. absorb. She chose the sponge for its . [absorption]
20. active. He tired after so much . [activity]
21. swim. She was a strong . [swimmer]
22. human. The kind man was known for his . [humanity]
23. wash. Put the laundry in the . [washer]
24. humor. The story was quite . [humorous]
25. assist. The teacher will give you . [assistance]
26. mystery. The dark glasses made the man look . [mysterious]
27. produce. The play was a grand . [production]
28. glory. The view from the hill top was . [glorious]
188 JOANNE F. CARLISLE

Part 2: Decomposition

Practice: a. Driver. Children are too young to . [drive]


b. Improvement. My teacher wants my spelling to . [improve]
1. growth. She wanted her plant to . [grow]
2. dryer. Put the wash out to . [dry]
3. variable. The time of his arrival did not . [vary]
4. width. The mouth of the river is very . [wide]
5. density. The smoke in the room was very . [dense]
6. discussion. The friends have a lot to . [discuss]
7. famous. The actor would achieve much . [fame]
8. description. The picture is hard to . [describe]
9. fifth. The boy counted from one to . [five]
10. election. Which person did they ? [elect]
11. strength. The girl was very . [strong]
12. decision. The boy found it hard to . [decide]
13. popularity. The girl wants to be . [popular]
14. runner. How fast can she ? [run]
15. publicity. His views were made . [public]
16. difference. Do their opinions ? [differ]
17. originality. That painting is very . [original]
18. agreeable. With that statement I could not . [agree]
19. courageous. The man showed great . [courage]
20. admission. How many people will they ? [admit]
21. dangerous. Are the children in any ? [danger]
22. reduction. The overweight man was trying to . [reduce]
23. baker. She put the bread in to . [bake]
24. division. The cake is hard to . [divide]
25. guidance. The map was her . [guide]
26. continuous. How long will the storm ? [continue]
27. reliable. On his friend he could always . [rely]
28. acceptance. Is that an offer you can ? [accept]
AWARENESS OF STRUCTURE AND MEANING 189

Appendix 2. Word reading test

Set A Set B
Transparent Shift puzzlement
powerful explanation* secretive
suddenly easily* cookery
harmful solution* corrective
movement curiosity* pailful
addition natural* organist
friendly heavily* equalize
government explosion* wifelike
lovely trial** sparkly
quickly daily** oddity
moisture*** preventive
combination* odorous
shiny** queendom
invention*** stardom
direction*** dramatize
definition* fearsome
beautiful* idealize
bucketful
flowery
beastly

* Words with both phonological and orthographic shifts.


** Words with orthographic shifts.
*** Words with phonological shifts.

References
Anglin, J.M. (1993). Vocabulary development: A morphological analysis, Monographs of the
Society for Research in Child Development 58, Serial #238.
Brittain, M.M. (1970). Inflectional performance and early reading achievement, Reading
Research Quarterly 6: 34–48.
Carlisle, J.F. (1988). Knowledge of derivational morphology and spelling ability in fourth,
sixth, and eighth graders, Applied Psycholinguistics 9: 247–266.
Carlisle, J.F. (1995). Morphological awareness and early reading achievement. In: L.B.
Feldman (ed.), Morphological aspects of language processing (pp. 189–209). Hillsdale,
NJ: Erlbaum.
Carlisle, J.F. & Nomanbhoy, D. (1993). Phonological and morphological development,
Applied Psycholinguistics 14: 177–195.
Carroll, J.B., Davies, P. & Richman, B. (1971). Word frequency book. New York: American
Heritage Publication Co.
Champion, A. (1997). Knowledge of suffixed words: A comparison of reading disabled and
nondisabled readers, Annals of Dyslexia 47: 29–55.
190 JOANNE F. CARLISLE

Chialant, D. & Caramazza, A. (1995). Where is morphology and how is it processed? In: L.
Feldman (ed.), Morphological aspects of language processing (pp. 55–76). Hillsdale, NJ:
Erlbaum.
Chomsky, C. (1970). Reading, writing, and phonology, Harvard Educational Review 40: 287–
309.
Cole, P., Beauvillain, C. & Segui, J. (1989). On the representation and processing of prefixed
and suffixed derived words: A differential frequency effect, Journal of Memory and
Language 28: 1–13.
Derwing, B.L. (1976). Morpheme recognition and the learning of rules for derivational
morphology, Canadian Journal of Linguistics 21: 38–66.
Elbro, C. & Arnbak, E. (1996). The role of morpheme recognition and morphological
awareness in dyslexia, Annals of Dyslexia 46: 209–240.
Fowler, A.E. & Liberman, I.Y. (1995). The role of phonology and orthography in morpholo-
gical awareness. In: L.B. Feldman (ed.), Morphological aspects of language processing
(pp. 157–188). Hillsdale, NJ: Erlbaum.
Freyd, P. & Baron, J. (1982). Individual differences in acquisition of derivational morphology,
Journal of Verbal Learning and Verbal Behavior 21: 282–295.
Jones, N.K. (1991). Development of morphophonemic segments in children’s mental repres-
entations of words, Applied Psycholinguistics 12: 217–239.
Leong, C.K. (1989). Productive knowledge of derivational rules in poor reader, Annals of
Dyslexia 39: 94–115.
Litowitz, B. (1976). Learning to make definitions, Journal of Child Language 4: 289–304.
Mahoney, D. (1994). Using sensitivity to word structure to explain variance in high school and
college level reading ability, Reading and Writing: An Interdisciplinary Journal 6: 19–44.
Nagy, W.E. & Anderson, R. (1984). The number of words in printed school English, Reading
Research Quarterly 19: 304–330.
Napps, S.E. (1989). Morphemic relationships in the lexicon: Are they distinct from semantic
and formal relationships?, Memory and Cognition 17: 729–739.
Schreuder, R. & Baayen, R.H. (1995). Modeling morphological processing. In: L.B. Feldman
(ed.), Morphological aspects of language processing (pp. 131–154). Hillsdale, NJ:
Erlbaum.
Snow, C.E. (1990). The development of definitional skill, Journal of Child Language 17: 697–
710.
Templeton, S. & Scarborough-Franks, L. (1985). The spelling’s the thing: Knowledge of
derivational morphology in orthography and phonology among older students, Applied
Psycholinguistics 6: 371–390.
Tyler, A. & Nagy, W. (1989). The acquisition of English derivational morphology, Journal of
Memory and Language 28: 649–667.
Tyler, A. & Nagy, W. (1990). Use of derivational morphology during reading, Cognition 36:
17–34.
Windsor, J. (1994). Children’s comprehension and production of derivational suffixes, Journal
of Speech and Hearing Research 37: 408–417.
Wysocki, K. & Jenkins, J.R. (1987). Deriving word meanings through morphological gener-
alization, Reading Research Quarterly 22: 66–81.

Address for correspondence: Professor Joanne Carlislie, Department of Communication,


Sciences and Disorders, Northwestern University, 2299 North Campus Drive, Evanston, IL,
USA
Phone: (847) 491-2497; Fax: (847) 491-2494; E-mail: j-carlislie@nwu.edu

You might also like