Академический Документы
Профессиональный Документы
Культура Документы
Bertram, R., Laine, M. & Virkkala, M. M. (2000). The role of derivational morphology in vocabulary acquisition: Get by with a little help
from my morpheme friends. Scandinavian Journal of Psychology, 41, 287296.
This study explores the role of morphology in vocabulary knowledge of 3 rd and 6 th grade Finnish elementary school children. In a word
definition task, children from both grades performed overall better on derived words than on monomorphemic words. However, the results
were modified by the factors Frequency and Productivity. Most strikingly, performance on monomorphemic words was disproportionately
weaker than on derived words at the low frequency range. At the high-frequency range, derived words with low-productive suffixes yielded
poorest performance. We partly make an appeal to the lexical-statistical properties of the Finnish language to explain the interaction of
Frequency and Word Structure. At any rate, the results suggest that Finnish elementary school children benefit significantly from utilizing
morphology in determining word meanings.
Key words: Morphology, acquisition, Finnish, lexicon, productivity, frequency.
Raymond Bertram, Department of Psychology, University of Turku, FIN-20520 Turku, Finland. E-mail: rayber@utu.fi
# 2000 The Scandinavian Psychological Associations. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and
350 Main Street, Malden, MA 02148, USA. ISSN 0036-5564.
288
289
EXPERIMENT 1
Method
290
Root
frequency
Lemma
frequency
Surface
frequency
Bigram
frequency
Word
length
in letters
Number
of items
Suffixes
employed
Function
Example
High-frequent derivation
with a high-productive suffix
314
32.8
6.4
1094
8.00
15
High-frequent derivation
with a low-productive suffix
181
33.9
8.4
1181
7.73
15
38.2
6.5
1120
7.20
-jA
-Us
-ntA
-stO
-mO
-lA
none
laula ja ``singer''
ilmoit us ``announcement''
kerro nta ``narration''
laiva sto ``fleet''
kampaa mo ``hairdresser's''
kahvi la ``coffee bar''
kellari ``cellar''
-jA
-Us
-ntA
-stO
-mO
-lA
none
somista ja ``decorator''
kuitta us ``receipt''
haudo nta ``bathing''
taru sto ``mythology''
sulatta mo ``meltery''
juusto la ``cheesery''
sikerma ``cluster''
Condition
High-frequent
monomorphemic word
Low-frequent derivation
with a high-productive suffix
Low-frequent derivation
with a low-productive suffix
Low-frequent
monomorphemic word
29.7
0.81
0.20
1359
8.27
15
28.5
1.00
0.23
1297
8.13
15
0.86
0.22
1163
7.20
RESULTS
Prior to statistical analysis we excluded two opaque items
from the low-frequent low-productive derivations, for
which according to three Finnish native speakers it
was impossible to calculate the meaning via the constituent
morphemes. All other items were included as they were
within two standard deviations of their condition mean.
Two children were discarded due to overall non-responsiveness.
frequency range (0.67 for high-productive and 0.49 for lowproductive derived words). Separate ANOVA's at the high
and low frequency range both showed a significant effect for
morphological structure (F1(2, 58) 42.2, p < 0.001 and
F1(2, 58) 44.2, p < 0.001, respectively). Post-hoc comparisons revealed that all contrasts differed significantly from
each other (all p's < 0.05). This means that at the highfrequency range high-productive derivations elicit the best
performance, followed by monomorphemic words and lowproductive derivations. At the low-frequency range, again
high-productive derivations yield best performance, followed by low-productive derivations and by monomorphemic words.
If we look at the distribution of 0-, 1-, and 2-point
definitions, it appears that the percentage of 0-point
definitions is higher for the monomorphemic words than
for the derived items (monomorphemic words 33.3%, lowproductive derivations 14.8%, high-productive derivations
16.1%). For the 2-point definitions, the proportion is
greater for high-productive derivations than for the other
two word types (monomorphemic words 48.3%, lowproductive derivations 46.3%, high-productive derivations
66.7%), yielding a highly significant effect ( 2(4)
164.9, p < 0.001, with Yates' continuity correction).
Scoring on the basis of the suffix for zero point responses
For all the 130 instances where an attempt to define a
derived word yielded zero points, it was decided whether
suffix knowledge was present in the answer (either formally
or semantically). In 71.5% of all these instances (N 93),
this was indeed the case. Most interestingly, a chi-square
shows that suffix-related knowledge is not randomly
distributed over the high- and low-productive condition
( 2(1) 22.8, p < 0.001, with continuity correction). Suffix
knowledge was much more common on the high-productive
condition (60 out of 66) than success on the low-productive
condition (33 out of 64).
DISCUSSION OF EXPERIMENT 1
In line with previous research, this study shows a robust
frequency effect. Children's definitions of high-frequency
words were in ordinal scale nearly twice as good as
Table 2. Mean definition scores with SD of the 6 target conditions of the 3 rd graders in experiment 1
Morphological structure
Mean scores
High-frequent words
1.84 (0.19)
1.54 (0.33)
1.70 (0.36)
Low-frequent words
1.17 (0.52)
1.05 (0.56)
0.60 (0.29)
Average
1.51
1.32
1.15
Average
1.69
0.94
1.33
291
292
EXPERIMENT 2
METHOD
Participants Thirty-two elementary school children (on the average
12 : 6 years of age) from two 6 th grade classes of the Puolala School
in Turku, Finland, were tested individually in a quiet room on
location. All of them were native speakers of Finnish.
Materials Seventy target words were selected from our lexical
database, 35 representing the high-frequency range and 35 the lowfrequency range. The high-frequency words were exactly the same as
the ones employed for the 3 rd grade, but about 43% of the lowfrequency words were replaced in order to increase the degree of
difficulty. As in Experiment 1, both frequency conditions included
three different word types: derived words with a high-productive
suffix, derived words with a low-productive suffix and monomorphemic words. In Table 3, the relevant quantitative data of the three
conditions at the low-frequency range (the ones that differed from
Experiment 1) are presented. As in Experiment 1, word length in
letters, lemma frequency, surface frequency and bigram frequency
were controlled. The root frequency difference between the highproductive derivations (7.1) and the low-productive derivations
(18.9) almost reached significance (t2(28) 1.94, p2 0.06). We
discuss the implications for this root frequency bias after presentation of the results.
Condition
Low-frequent derivation
with a high-productive suffix
Low-frequent derivation
with a low-productive suffix
Low-frequent
monomorphemic word
Lemma
frequency
Surface
frequency
Bigram
frequency
Word
length
in letters
Number
of items
Suffixes
employed
7.1
0.80
0.16
1100
7.80
15
18.9
0.52
0.15
1237
7.93
15
1.05
0.28
1361
8.20
-jA, -Us,
-ntA
-stO, -mO,
-lA
none
Root
frequency
Example
kuitta us
``receipt''
taru sto
``mythology''
sikerma
``cluster''
RESULTS
Prior to statistical analyses we excluded four opaque items
from the low-frequent low-productive derivations. All other
items were included as they were within two standard
deviations of their condition mean. All children were
included in the analyses as well.
Scoring on the basis of the whole word form
The responses of the remaining 66 definitions of the 32
children were used to calculate the mean scores per
condition. They are presented in Table 4.
A 2 3 repeated measures ANOVA revealed a significant
main effect for Frequency (F1(1, 31) 192.8, p < 0.001),
Morphological Structure (F1(2, 62) 23.6, p < 0.001), as well
as a significant interaction (F1(2, 62) 52.2, p < 0.001).
Again, the high-frequent items were better defined than the
low-frequent items. With respect to the main effect of
Morphological Structure, post-hoc comparisons showed that
the words with the high-productive suffixes were better
defined than the words with the low-productive suffixes
(F1(1, 29) 9.0, p < 0.001) or the monomorphemic words
(F1(1, 29) 36.7, p < 0.001). Moreover, the difference between words with low-productive suffixes and monomorphemic words was significant (F1(1, 29) 16.4, p < 0.001), the
former being defined more accurately than the latter. As in
Experiment 1, the significant interaction was accounted for
mainly by the large performance difference between the highvs. low-frequent monomorphemic words (a difference of 1.03
points), whereas performance on both types of derived words
drops down less drastically when moving from high- to lowfrequency range (0.69 for high-productive and 0.41 for lowproductive derived words). Separate ANOVA's at the high
and low frequency range both showed once again a significant
effect for morphological structure (F1(2, 62) 42.4, p < 0.001
and F1(2, 58) 37.1, p < 0.001, respectively). Post-hoc comparisons revealed that all contrasts but one significantly differ
from one another ( p's < 0.05). The only exception was the
contrast between high-productive derivations and the monomorphemic words at the high-frequency range ( p > 0.15).
This means that at the high-frequency range, high-productive
derivations and monomorphemic words elicit the best
performance, followed by low-productive derivations. At
the low-frequency range, the low-productive derivations yield
best performance, followed by the high-productive derivations and by the monomorphemic words.
If we look at the distribution of 0-, 1-, and 2-point
definitions, we see again that for monomorphemic items the
percentage of 0-point whole word definitions higher than for
the derived words (monomorphemic words 26.1%, lowproductive derivations 7.4%, high-productive derivations
14.4%). For the 2-point definitions, the proportion is
greater for high-productive derivations than for the other
two word types (monomorphemic words 65.4%, lowproductive derivations 64.6%, high-productive derivations
73.7%), yielding a highly significant difference
( 2(4) 148.0, p < 0.001, with Yates' continuity correction).
Scoring on the basis of the suffix for zero point responses
Of all the 122 instances where an attempt to define a derived
word yielded zero points, 75.4% (N 92) scored one point
for the suffix. In contrast with Experiment 1, a chi-square
test on the distribution of the correctly scored items shows
that they are randomly distributed over the high- and lowproductive condition ( 2(1) 1.83, p > 0.1, with continuity
correction; for the high-productive condition the score was
55 out of 68, for the low-productive condition 37 out of 54).
DISCUSSION OF EXPERIMENT 2
The pattern of results found for the 6 th graders mirrors that
of the 3 rd graders in many respects. Overall, performance for
complex words is better than that for simple words, which is
caused chiefly by the large differences at the low-frequency
range. Moreover, overall scores indicate better performance
for high-productive than for low-productive derived words.
However, when looking at the frequency ranges separately,
two noticeable differences with the 3 rd graders data pattern
are observable.
First, at the high-frequency range monomorphemic words
and high-productive derivations are now yielding similar
scores. We believe that this is due to a ceiling effect, both
categories scoring close to maximum. Second, at the lowfrequency range low-productive derivations surprisingly
enough elicit best performance. This could be partly
ascribed to a growing morphological knowledge in general
and for low-productive suffixes in particular, also supported
Table 4. Mean definition scores with SD of the 6 target conditions of the 6 th graders in experiment 2
Morphological structure
Mean scores
High-frequent words
1.94 (0.07)
1.74 (0.21)
1.91 (0.10)
Low-frequent words
1.25 (0.64)
1.33 (0.50)
0.88 (0.62)
Average
1.59
1.54
1.39
Average
1.86
1.15
1.51
293
294
GENERAL DISCUSSION
Morphological knowledge matters. As already shown by
previous studies (most notably, Anglin, 1993, and Nagy &
Anderson, 1984), children acquire a more extensive vocabulary by making use of their ability to analyze and
comprehend words via morphological constituents. The
general picture that arises from this study is that on top of
that they have a greater understanding of morphologically
complex than simple words when the word types are tightly
matched on all relevant factors. However, we were able to
acquire a more detailed picture by manipulating factors such
as frequency of occurrence and suffix productivity in a
relatively unexplored but from a morphologically point-ofview very interesting language, namely Finnish.
Suffix productivity turned out to be an important
differentiating factor. High-productive derivations were
understood best by children of both the 3 rd and the 6 th
grade, although in the latter grade performance on lowproductive derivations seemed to approach that on highproductive ones. Why would suffix productivity be so
important in vocabulary acquisition? A reasonable answer
on this question could be given by referring to what Tyler
and Nagy (1989) call relational knowledge. They define this
type of knowledge as the ability to recognize morphological
relations between words that share common morphemes
such as work and worker, or for that matter, worker and
thinker. It is not hard to see that a morphological pattern
that gets reinstated regularly in many word formations at
both input and output will be acquired earlier and more
strongly than a pattern that is encountered or produced less
often. In general, it becomes more and more apparent that
affixal properties have a huge impact on processing of
morphologically complex words (e.g., Bertram et al., 1999;
Bertram et al., 2000b; Burani, Dovetto, Thornton, &
Laudanna, 1997; Laudanna & Burani, 1995; Schreuder &
Baayen, 1995).
# 2000 The Scandinavian Psychological Associations.
NOTES
1
There are exceptions to this observation. Lewis and Windsor
(1996), for instance, assessed the effects of suffix productivity on
nonsense derivations. The more productive the suffix, the better
production and comprehension performance on the nonsense
derivations was observed.
2
We define a morphological family as the group of derived and
compound words that are sharing the same root. Thus words like
``worker'', ``work-man'', ``unworkable'' belong to the same morphological family. In a recent article Bertram, Baayen, & Schreuder
(2000a) assess the effects of morphological family size on lexical
processing.
3
The database mentioned here has compiled words of articles of
the Turun Sanomat, the second largest newspaper in Finland. The
compilation stretches from 1.4.1994 to 30.6.1996. There are 22.7
million word forms divided over 1, 483, 912 distinct word types.
Nouns are clearly the largest grammatical class in this database; the
two other major classes, verbs and adjectives, account for 118, 521
and 219, 984.
4
This percentage is actually even lower, for the low-productive
derived word types are still included in the count of 26, 355, since
they are not recognized by the automatic morphological parser
employed in the tagging of our lexical database.
5
In order to assess the degree of productivity for certain
derivational affixes of Finnish, a production experiment with 37
adult participants was conducted in which the participants had to
create as many words as possible with a given suffix in a limited
amount of time. For the high-productive suffixes -jA, -Us, and
-ntA, the growth rate (see Baayen, 1994) was 0.322, 0.491, and
0.289, respectively. For the low-productive suffixes -mO, -stO,
and -lA, the growth rate was 0.275, 0.201, and 0.113, respectively.
6
A two-tailed t-test for independent samples revealed that the
root frequency of the two derivation conditions was matched indeed
at both frequency ranges (low-frequency range t2(28) < 1; highfrequency range t2(28) 1.05, p2 < 0.30), even though the absolute
difference in the latter range seems quite large (181 vs. 314). This
difference, however, is caused by one outlier in the high-productive
derivation condition with a root frequency of 1638. Without this
item the average root frequency for high-productive derivations
would have been 219.
REFERENCES
Anglin, J. M. (1993). Vocabulary development: a morphological
analysis. Monographs of the Society for Research in Child
Development. (Serial No. 238). Chicago: University of Chicago
Press.
Baayen, R. H. (1994). Productivity in language production.
Language and Cognitive Processes, 9, 447469.
Baayen, R. H., Dijkstra, T. & Schreuder, R. (1997). Singulars and
plurals in Dutch: Evidence for a parallel dual route model.
Journal of Memory and Language, 36, 94 117.
Bertram, R., Baayen, R. H. & Schreuder, R. (2000a). Effects of
family size for complex words. Journal of Memory and
Language, 42, 390405.
Bertram, R., Laine, M. & Karvinen, K. (1999). The interplay of
Word Formation Type, Affixal Homonymy, and Productivity in
lexical processing: evidence from a morphologically rich
language. Journal of Psycholinguistic Research, 28, 213 226.
Bertram, R., Schreuder, R. & Baayen, R. H. (2000b). The balance of
# 2000 The Scandinavian Psychological Associations.
295
296