L2 Vocabulary Acquisition in Children: Effects of Learning Method and Cognate Status

Language Learning ISSN 0023-8333
L2 Vocabulary Acquisition in Children:
Effects of Learning Method and Cognate
Status
Claudio Tonzar
University of Urbino
Lorella Lotto
University of Padova
Remo Job
University of Trento, Italy
In this study we investigated the effects of two learning methods (picture- or wordmediated
learning) and of word status (cognates vs. noncognates) on the vocabulary
acquisition of two foreign languages: English and German.We examined children from
fourth and eig...

At testing, participants were presented
with both congruent (i.e., words for the word-learning group) and incongruent
(i.e., pictures for the word-learning group) stimuli.

At testing, participants were presented
with both congruent (i.e., words for the word-learning group) and incongruent
(i.e., pictures for the word-learning group) stimuli.

Fourth-Grade Students
The main effects and interactions that proved significant for either English
and/or German are shown in Table 2.

Both models assumed a direct link between the lexicon of the L1 and
the conceptual system.

Language Learning ISSN 0023-8333
L2 Vocabulary Acquisition in Children:
Effects of Learning Method and Cognate
Status
Claudio Tonzar
University of Urbino
Lorella Lotto
University of Padova
Remo Job
University of Trento, Italy
In this study we investigated the effects of two learning methods (picture- or wordmediated
learning) and of word status (cognates vs. noncognates) on the vocabulary
acquisition of two foreign languages: English and German.We examined children from
fourth and eig...

Furthermore, in order to assess the
efficacy of the two methods over time, participants’ performance was evaluated
in four different experimental sessions.

Language Learning ISSN 0023-8333
L2 Vocabulary Acquisition in Children:
Effects of Learning Method and Cognate
Status
Claudio Tonzar
University of Urbino
Lorella Lotto
University of Padova
Remo Job
University of Trento, Italy
In this study we investigated the effects of two learning methods (picture- or wordmediated
learning) and of word status (cognates vs. noncognates) on the vocabulary
acquisition of two foreign languages: English and German.We examined children from
fourth and eig...

Language Learning ISSN 0023-8333
L2 Vocabulary Acquisition in Children:
Effects of Learning Method and Cognate
Status
Claudio Tonzar
University of Urbino
Lorella Lotto
University of Padova
Remo Job
University of Trento, Italy
In this study we investigated the effects of two learning methods (picture- or wordmediated
learning) and of word status (cognates vs. noncognates) on the vocabulary
acquisition of two foreign languages: English and German.We examined children from
fourth and eig...

As a consequence, it may
be plausible to hypothesize that in learning a new language, setting up a new
entry in memory is more demanding than modifying a preexisting similar
entry, and this would facilitate the learning of L2 cognates over noncognates.

This pattern
of a stronger superiority of the picture-learning method in the latter sessions
shows that the effects of the picture-mediated learning method may be fully
appreciated when the learning sessions are spaced in time.

After a learning phase during which L2
words were presented either together with the corresponding translation in L1 (wordlearning
method) or with the picture of the object to which the word refers (picturelearning
method), children were tested at different points in time.

Furthermore, at the methodological level, this allowed us
to check the performance on the same set of items, minimizing effects due to
specific item in one language.

Language Learning ISSN 0023-8333
L2 Vocabulary Acquisition in Children:
Effects of Learning Method and Cognate
Status
Claudio Tonzar
University of Urbino
Lorella Lotto
University of Padova
Remo Job
University of Trento, Italy
In this study we investigated the effects of two learning methods (picture- or wordmediated
learning) and of word status (cognates vs. noncognates) on the vocabulary
acquisition of two foreign languages: English and German.We examined children from
fourth and eighth ...

Both models assumed a direct link between the lexicon of the L1 and
the conceptual system.

Some of them
concern linguistic aspects (e.g., word class, degree of similarity between languages),
others concern the teaching methods adopted and the corresponding
learning strategies, still others relate to the learner (e.g., age, motivation, knowledge
of other languages), and so on.

Although both models describe
the processing of L2 words, it has been proposed that they refer to different
levels of proficiency in L2. Specifically, Kroll and Curley (1988) and Chen
and Leung (1989) tested adult beginners with different levels of expertise in
the L2, showing that whereas the more proficient individuals were equally
fast in translating and naming pictures in L2, the less proficient ones were
faster in translating than in naming pictures in the L2. Thus, the performan...

Category interference in translation and picture
naming: Evidence for asymmetric connections between bilingual memory
representations.

In our study, the cognate stimuli were cognates in all three languages, and this
would maximize the possibility of detecting a difference between cognates and
noncognate stimuli.

After a learning phase during which L2
words were presented either together with the corresponding translation in L1 (wordlearning
method) or with the picture of the object to which the word refers (picturelearning
method), children were tested at different points in time.

Specifically, we (a) compared two teaching/learning methods
in four experimental sessions with different temporal features, (b) evaluated the
vocabulary acquisition of two foreign languages by learners of two age groups,
and (c) assessed the role of words’ cognate status in the learning process.

The results showed
that the picture-based method leads to a better performance than the word-basedmethod,
but this effect was modulated by cognate status and age of learning.

In a rendition
of an English or a German word, two (possible) ways to maximize similarities
between the two languages are the following:
1. transforming L2 phonemes into “canonical” L1 graphemes (e.g., writing
“Carafe” for “Karaffe,” because in Italian the sound /k/ is spelled as /c/);
2. adapting L2 words to L1 word forms (e.g., adding a vowel in last position:
“Polipe” instead of “Polyp,” because nouns in Italian tend to end with a
vowel).

On the one hand,
cognates were found to facilitate the acquisition of L2 words, but when the language was
familiar (i.e.,

Specifically, we (a) compared two teaching/learning methods
in four experimental sessions with different temporal features, (b) evaluated the
vocabulary acquisition of two foreign languages by learners of two age groups,
and (c) assessed the role of words’ cognate status in the learning process.

This pattern
of a stronger superiority of the picture-learning method in the latter sessions
shows that the effects of the picture-mediated learning method may be fully
appreciated when the learning sessions are spaced in time.

On the other hand, cognate status interacted with the learning method in the
older children, indicating that the word method was particularly effective for cognate
words.

In R. J.
Harris (Ed.), Cognitive processing in bilinguals (pp. 253–264).

The results showed a
significant congruency effect: Participants were more efficient with congruent
stimuli, in other words, the word-learning group was faster in translating in
L2, whereas the picture-learning group was faster in naming pictures in L2.
However, a main effect of learning method was not found.

Memory and Cognition, 18, 279–288.

Language Learning 59:3, September 2009, pp. 623–646 624
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
The results of the few studies that have focused on the role of these two
different methods of vocabulary acquisition challenge current models of lexical
representation in bilinguals.

The more often cited study on this issue was
conducted by Potter, So, von Eckhart, and Feldman (1984), who tested two
models of word processing and representation of the corresponding linguistic
knowledge in memory: the word association model and the concept mediation
model.

Language Learning ISSN 0023-8333
L2 Vocabulary Acquisition in Children:
Effects of Learning Method and Cognate
Status
Claudio Tonzar
University of Urbino
Lorella Lotto
University of Padova
Remo Job
University of Trento, Italy
In this study we investigated the effects of two learning methods (picture- or wordmediated
learning) and of word status (cognates vs. noncognates) on the vocabulary
acquisition of two foreign languages: English and German.We examined children from
fourth and eig...

The authors
proposed that young beginners used pictorial representations rather than L1
words as media to produce the corresponding L2 responses.

The more often cited study on this issue was
conducted by Potter, So, von Eckhart, and Feldman (1984), who tested two
models of word processing and representation of the corresponding linguistic
knowledge in memory: the word association model and the concept mediation
model.

However, this concern may be
attenuated if we consider that mean word length was controlled for between
languages and that the same set of stimuli was used to compare the two methods.

Specifically, we (a) compared two teaching/learning methods
in four experimental sessions with different temporal features, (b) evaluated the
vocabulary acquisition of two foreign languages by learners of two age groups,
and (c) assessed the role of words’ cognate status in the learning process.

Lexical processing in bilingual or multilingual speakers.

Specifically, we (a) compared two teaching/learning methods
in four experimental sessions with different temporal features, (b) evaluated the
vocabulary acquisition of two foreign languages by learners of two age groups,
and (c) assessed the role of words’ cognate status in the learning process.

Language Learning ISSN 0023-8333
L2 Vocabulary Acquisition in Children:
Effects of Learning Method and Cognate
Status
Claudio Tonzar
University of Urbino
Lorella Lotto
University of Padova
Remo Job
University of Trento, Italy
In this study we investigated the effects of two learning methods (picture- or wordmediated
learning) and of word status (cognates vs. noncognates) on the vocabulary
acquisition of two foreign languages: English and German.We examined children from
fourth and eig...

The results showed a
significant congruency effect: Participants were more efficient with congruent
stimuli, in other words, the word-learning group was faster in translating in
L2, whereas the picture-learning group was faster in naming pictures in L2.
However, a main effect of learning method was not found.

Both models assumed a direct link between the lexicon of the L1 and
the conceptual system.

For each name, cognate status ratings,
length, name agreement, and word frequency are shown in the Appendix.1

Some of them
concern linguistic aspects (e.g., word class, degree of similarity between languages),
others concern the teaching methods adopted and the corresponding
learning strategies, still others relate to the learner (e.g., age, motivation, knowledge
of other languages), and so on.

In that study, a similar methodology and a similar
population were used; that is, a verbal method and a visual method were used
to assess the acquisition of L2 words in fourth-grade students.

Language Learning 59:3, September 2009, pp. 623–646 623
C 
2009 Language Learning Research Club, University of Michigan
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
Introduction
Several factors affect the learning of a second language (L2).

Keywords second language learning; L2 vocabulary learning; cognate status; teaching/
learning methods; immediate and delayed learning; concurrent acquisition of two foreign
languages
Correspondence concerning this article should be addressed toClaudio Tonzar, Institute of Psychology,
University of Urbino, Via Ubaldini 17, 61029 Urbino (PU), Italy.

As expected, there was an effect
of cognate status, with cognates being processed more accurately than noncognates.

Some of them
concern linguistic aspects (e.g., word class, degree of similarity between languages),
others concern the teaching methods adopted and the corresponding
learning strategies, still others relate to the learner (e.g., age, motivation, knowledge
of other languages), and so on.

The teaching/learning method may play a relevant role in the acquisition of
a new language because it can modulate lexical processing.

Results
In Figure 1, the percentages of correct responses are plotted as a function of
learning method, cognate status, language, and grade.

After a learning phase during which L2
words were presented either together with the corresponding translation in L1 (wordlearning
method) or with the picture of the object to which the word refers (picturelearning
method), children were tested at different points in time.

In A. M. B. De Groot & J. F.
Kroll (Eds.), Tutorials in bilingualism: Psycholinguistic perspectives
(pp. 169–199).

Two ways to maximize differences are the following:
3. using letters of L2 that are not part of the L1 alphabet (e.g., inserting “w,”
“y,”or “ß,” which are not part of the Italian alphabet; “Flatoß” instead of
“Fl¨ote”);
4. adapting L1 words to L2 word forms (e.g., deleting a vowel, or adding a
consonant, thus differentiating it from Italian words that generally have a
final-position vowel; “M¨uhler” instead of M¨uhle).

This pattern of results appeared to be very robust (for reviews, see Chen,
1992; De Groot, 1993; Kroll, 1993; Kroll & De Groot, 1997; Kroll, Michael,
& Sankaranarayanan, 1998; Kroll & Tokowicz, 2001) and led Kroll and Stewart
(1994) to propose the revised hierarchical model (RHM), which assumes
that less fluent bilinguals process words in the L2 by relying on their lexical
knowledge in the L1, whereas more proficient bilinguals use the concept-to-L2
link to process words without activatin...

This is a quite common situation
Language Learning 59:3, September 2009, pp. 623–646 638
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
that has attracted quite a lot of attention from the general public but about which
not much rigorous empirical evidence is available at present.

Keywords second language learning; L2 vocabulary learning; cognate status; teaching/
learning methods; immediate and delayed learning; concurrent acquisition of two foreign
languages
Correspondence concerning this article should be addressed toClaudio Tonzar, Institute of Psychology,
University of Urbino, Via Ubaldini 17, 61029 Urbino (PU), Italy.

Furthermore, in order to assess the
efficacy of the two methods over time, participants’ performance was evaluated
in four different experimental sessions.

In a rendition
of an English or a German word, two (possible) ways to maximize similarities
between the two languages are the following:
1. transforming L2 phonemes into “canonical” L1 graphemes (e.g., writing
“Carafe” for “Karaffe,” because in Italian the sound /k/ is spelled as /c/);
2. adapting L2 words to L1 word forms (e.g., adding a vowel in last position:
“Polipe” instead of “Polyp,” because nouns in Italian tend to end with a
vowel).

The more often cited study on this issue was
conducted by Potter, So, von Eckhart, and Feldman (1984), who tested two
models of word processing and representation of the corresponding linguistic
knowledge in memory: the word association model and the concept mediation
model.

This pattern of results appeared to be very robust (for reviews, see Chen,
1992; De Groot, 1993; Kroll, 1993; Kroll & De Groot, 1997; Kroll, Michael,
& Sankaranarayanan, 1998; Kroll & Tokowicz, 2001) and led Kroll and Stewart
(1994) to propose the revised hierarchical model (RHM), which assumes
that less fluent bilinguals process words in the L2 by relying on their lexical
knowledge in the L1, whereas more proficient bilinguals use the concept-to-L2
link to process words without activ...

A further observation about the
relative “independence” of two languages learned concurrently involves the rate
of substitutions of the two languages: Very rarely words in one language were
substituted for by words in the other language.

In line with a developmental
perspective, we expected that, overall, older students performed better
than younger students.

This pattern of results appeared to be very robust (for reviews, see Chen,
1992; De Groot, 1993; Kroll, 1993; Kroll & De Groot, 1997; Kroll, Michael,
& Sankaranarayanan, 1998; Kroll & Tokowicz, 2001) and led Kroll and Stewart
(1994) to propose the revised hierarchical model (RHM), which assumes
that less fluent bilinguals process words in the L2 by relying on their lexical
knowledge in the L1, whereas more proficient bilinguals use the concept-to-L2
link to process words without activ...

For our
participants, such superiority tends to be smaller in the first testing session
and larger in the latter sessions although there is some variability probably
due to the very low performance for German in some conditions.

Students in neither grade
had been systematically exposed to German, although eighth graders had been
taught English in the previous 2 years, albeit on a very restricted schedule.

This pattern of results appeared to be very robust (for reviews, see Chen,
1992; De Groot, 1993; Kroll, 1993; Kroll & De Groot, 1997; Kroll, Michael,
& Sankaranarayanan, 1998; Kroll & Tokowicz, 2001) and led Kroll and Stewart
(1994) to propose the revised hierarchical model (RHM), which assumes
that less fluent bilinguals process words in the L2 by relying on their lexical
knowledge in the L1, whereas more proficient bilinguals use the concept-to-L2
link to process words without activ...

The issue
of learning two L2s at a time is a theoretically interesting matter because
it touches on the constraints of the acquisition mechanism involved in L2
learning.

The latter did not differ from the test following the
first learning session (.44).
631 Language Learning 59:3, September 2009, pp. 623–646
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
Table 3 Accuracy (proportion of correct responses) for the different experimental conditions
for English (fourth graders)
Test session
Method Cognate status 1st session 2nd session 1 week 1 month Mean
Picture-word Cognates .66 .80 .63 .59 .67

Specifically, we (a) compared two teaching/learning methods
in four experimental sessions with different temporal features, (b) evaluated the
vocabulary acquisition of two foreign languages by learners of two age groups,
and (c) assessed the role of words’ cognate status in the learning process.

In this article, we present a study aimed
at investigating how some of these factors affected vocabulary acquisition in a
school environment in which two foreign languages are acquired in close temporal
contiguity.

International Journal
of Bilingual Education and Bilingualism, 5, 267–278.

After a learning phase during which L2
words were presented either together with the corresponding translation in L1 (wordlearning
method) or with the picture of the object to which the word refers (picturelearning
method), children were tested at different points in time.

Although both models describe
the processing of L2 words, it has been proposed that they refer to different
levels of proficiency in L2. Specifically, Kroll and Curley (1988) and Chen
and Leung (1989) tested adult beginners with different levels of expertise in
the L2, showing that whereas the more proficient individuals were equally
fast in translating and naming pictures in L2, the less proficient ones were
faster in translating than in naming pictures in the L2. Thus, the performan...

Both models assumed a direct link between the lexicon of the L1 and
the conceptual system.

In line with a developmental
perspective, we expected that, overall, older students performed better
than younger students.

A word is considered a cognate when it is orthographically and/or
phonologically similar to its translation in another language.

The analyses of
variance (ANOVA) on proportion of correct responses were performed after a
2 arcsin transformation.

Summary
Let us briefly summarize the main results.

Specifically, we (a) compared two teaching/learning methods
in four experimental sessions with different temporal features, (b) evaluated the
vocabulary acquisition of two foreign languages by learners of two age groups,
and (c) assessed the role of words’ cognate status in the learning process.

The automatic cognate form assumption: Evidence for the parasitic
model of vocabulary development.

In addition, we should reiterate that
such superiority is context-sensitive.

Although both models describe
the processing of L2 words, it has been proposed that they refer to different
levels of proficiency in L2. Specifically, Kroll and Curley (1988) and Chen
and Leung (1989) tested adult beginners with different levels of expertise in
the L2, showing that whereas the more proficient individuals were equally
fast in translating and naming pictures in L2, the less proficient ones were
faster in translating than in naming pictures in the L2. Thus, the performan...

In line with a developmental
perspective, we expected that, overall, older students performed better
than younger students.

Language Learning 59:3, September 2009, pp. 623–646 630
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
Table 2 ANOVA summary table concerning fourth-grade participants
English German
Significant effects F p ηp
2 F p ηp
2
Learning method, F(1, 121) 9.94 .002 .08 13.09 .0004 .10

In the cognate-rating study, we presented 63 Italian-English, Italian-
German, and English-German word pairs corresponding to the names of the
pictures used in the first pretest and asked 20 third-year middle school Italian
children (13 years old) to evaluate the orthographic similarity within each word
pair on a 7-point scale.
627 Language Learning 59:3, September 2009, pp. 623–646
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
Table 1 Mean cognate ratings and mean length

This pattern of results highlights two aspects.

Furthermore, in order to assess the
efficacy of the two methods over time, participants’ performance was evaluated
in four different experimental sessions.

Participants
performed better in the picture-word than in the word-word condition (.55 vs.
.46), and they were more accurate with cognates than noncognates (averaging
across conditions, .64 vs. .36).

In this article, we present a study aimed
at investigating how some of these factors affected vocabulary acquisition in a
school environment in which two foreign languages are acquired in close temporal
contiguity.

Previous studies on L2
performance have shown that cognates are translated faster and more accurately
than noncognates, both in forward and backward direction (e.g.,

From a theoretical standpoint, this study provides evidence on the differential
role played by the links between the conceptual system and the L1 and L2
Table 8 Proportion of wrongly reported words for cognates and noncognates as a
function of the two strategies
Maximizing similarities Maximizing differences
Cognate Transform L2 into Adapt L2 to Insert peculiar Adapt L1 to
status “canonical” L1 L1 word forms L2 letters L2 word forms
Cognates 27.3 12.5 0.9 3.03

In fact, Lotto and De Groot (1998), although acknowledging
the possibility that the difference in processing cognate and non-cognate words
may be ascribed to the level of representation, pointed out that formal similarity
between cognate translations may act either at the learning level, as there is
less to learn with cognate words than with noncognate words, or at the retrieval
level, as cognates constitute a strong cue for the retrieval of the corresponding
translation.

In a rendition
of an English or a German word, two (possible) ways to maximize similarities
between the two languages are the following:
1. transforming L2 phonemes into “canonical” L1 graphemes (e.g., writing
“Carafe” for “Karaffe,” because in Italian the sound /k/ is spelled as /c/);
2. adapting L2 words to L1 word forms (e.g., adding a vowel in last position:
“Polipe” instead of “Polyp,” because nouns in Italian tend to end with a
vowel).

American Journal of Speech-Language Pathology, 11, 381–
394.

Two ways to maximize differences are the following:
3. using letters of L2 that are not part of the L1 alphabet (e.g., inserting “w,”
“y,”or “ß,” which are not part of the Italian alphabet; “Flatoß” instead of
“Fl¨ote”);
4. adapting L1 words to L2 word forms (e.g., deleting a vowel, or adding a
consonant, thus differentiating it from Italian words that generally have a
final-position vowel; “M¨uhler” instead of M¨uhle).

In that study, a similar methodology and a similar
population were used; that is, a verbal method and a visual method were used
to assess the acquisition of L2 words in fourth-grade students.

In the second
method (the word-word condition), L2 words were presented together with
the correspondingwords in the first language (L1).Afewstudies have addressed
this issue.

The three factors we manipulated—learning/teaching method, test session,
and cognate status—were all significant.

Language interference and language learning
techniques transfer in L2 and L3 immersion programmes.

In a rendition
of an English or a German word, two (possible) ways to maximize similarities
between the two languages are the following:
1. transforming L2 phonemes into “canonical” L1 graphemes (e.g., writing
“Carafe” for “Karaffe,” because in Italian the sound /k/ is spelled as /c/);
2. adapting L2 words to L1 word forms (e.g., adding a vowel in last position:
“Polipe” instead of “Polyp,” because nouns in Italian tend to end with a
vowel).

English for the eighth-grade children), this effect decreased with exposure
to language.

The finding that the word-learning method is more effective for cognate words
suggests that the formal similarity of words in the two languages may be a
helpful retrieval cue of the orthographic form of words in L2. This may be
particularly effective in a language like Italian, for which the mapping from
phonology to orthography is highly regular.

English for the eighth-grade children), this effect decreased with exposure
to language.

Immediate Test 1 was administered at the end of the first
learning session and Immediate Test 2 was administered at the end of the
second learning session.

The shift from one form of representation to the other is assumed
to be gradual and to be composed of changes in the relative weight given to
the lexical and conceptual links: As fluency increases, the existing L2-to-L1
lexical links are complemented, rather than substituted, by stronger additional
links between the L2 lexicon and the conceptual system.

This pattern of results appeared to be very robust (for reviews, see Chen,
1992; De Groot, 1993; Kroll, 1993; Kroll & De Groot, 1997; Kroll, Michael,
& Sankaranarayanan, 1998; Kroll & Tokowicz, 2001) and led Kroll and Stewart
(1994) to propose the revised hierarchical model (RHM), which assumes
that less fluent bilinguals process words in the L2 by relying on their lexical
knowledge in the L1, whereas more proficient bilinguals use the concept-to-L2
link to process words without activ...

On the one hand, because the formal similarity of
the two words may be a facilitating cue for retrieval, a larger effect of cognate
Language Learning 59:3, September 2009, pp. 623–646 626
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
status in theword-learning method than in the picture-learning method could be
predicted.

Some of them
concern linguistic aspects (e.g., word class, degree of similarity between languages),
others concern the teaching methods adopted and the corresponding
learning strategies, still others relate to the learner (e.g., age, motivation, knowledge
of other languages), and so on.

The issue
of learning two L2s at a time is a theoretically interesting matter because
it touches on the constraints of the acquisition mechanism involved in L2
learning.

Some implications concerning learning two languages at a time in the school
setting are discussed.

Some of them
concern linguistic aspects (e.g., word class, degree of similarity between languages),
others concern the teaching methods adopted and the corresponding
learning strategies, still others relate to the learner (e.g., age, motivation, knowledge
of other languages), and so on.

Some implications concerning learning two languages at a time in the school
setting are discussed.

A word is considered a cognate when it is orthographically and/or
phonologically similar to its translation in another language.

In A. F. Healy
& L. E. Bourn (Eds.), Foreign language learning: psycholinguistic experiment on
training and retention (pp. 365–395).

In this article, we present a study aimed
at investigating how some of these factors affected vocabulary acquisition in a
school environment in which two foreign languages are acquired in close temporal
contiguity.

In order to simplify the exposition of the results, the analyses for the two
age groups and the two languages will be presented separately.

Analyzing a subset of learners’ errors, we
found that when a cognate was wrongly reported, it conformed more often to
the strategy of maximizing the similarities, wherease when a noncognate was
wrongly reported, it conformed more often to the strategy of maximizing the
differences (see Table 8).

Although both models describe
the processing of L2 words, it has been proposed that they refer to different
levels of proficiency in L2. Specifically, Kroll and Curley (1988) and Chen
and Leung (1989) tested adult beginners with different levels of expertise in
the L2, showing that whereas the more proficient individuals were equally
fast in translating and naming pictures in L2, the less proficient ones were
faster in translating than in naming pictures in the L2. Thus, the performan...

The three 2-way interactions should be viewed in light of the triple interaction
Learning method × Cognate status × Test session.

If Italian children learning L2 exploited knowledge of their L1 when learning
in the L2, we should detect signs of interference from L1 to L2. For
example, words that are similar but have different meanings in two languages
(false friends) may hinder rather than help learning.

For instance, S´anchez-Casas et al. suggested
that cognates share the same lexical representation, whereas noncognates are
represented as separate entries in lexical memory.

This pattern of results appeared to be very robust (for reviews, see Chen,
1992; De Groot, 1993; Kroll, 1993; Kroll & De Groot, 1997; Kroll, Michael,
& Sankaranarayanan, 1998; Kroll & Tokowicz, 2001) and led Kroll and Stewart
(1994) to propose the revised hierarchical model (RHM), which assumes
that less fluent bilinguals process words in the L2 by relying on their lexical
knowledge in the L1, whereas more proficient bilinguals use the concept-to-L2
link to process words without activ...

Lexical memory in novice bilinguals: The role of
concepts in retrieving second language words.

Language Learning 59:3, September 2009, pp. 623–646 624
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
The results of the few studies that have focused on the role of these two
different methods of vocabulary acquisition challenge current models of lexical
representation in bilinguals.

Objects belonging to 12 semantic categories
(buildings, vehicles, weapons,musical instruments, animals, plants, fish,
kitchen utensils, fruits, tools, clothes and vegetables) were depicted.

The authors
proposed that young beginners used pictorial representations rather than L1
words as media to produce the corresponding L2 responses.

Objects belonging to 12 semantic categories
(buildings, vehicles, weapons,musical instruments, animals, plants, fish,
kitchen utensils, fruits, tools, clothes and vegetables) were depicted.

The present study may help to clarify this issue by examining
fourth- and eighth-grade children in a school setting.

The same two learning methods were contrasted by Lotto and De Groot
(1998), who instructed two groups of Dutch undergraduates to learn words in
a novel language (Italian).

Furthermore, at the methodological level, this allowed us
to check the performance on the same set of items, minimizing effects due to
specific item in one language.

Furthermore, it is relevant from a social point of view, as many
curricula provide the opportunity (or, sometimes, requirement) to learn two
languages at school, and yet no firm evidence exists on how this may affect the
outcome of the learning process.

The superiority showed by the picture method compared to the verbal
one over time suggests that the link among the structural description of an
object, its visual representation, and its name is more efficient than the link
between the corresponding L1 and L2 names.

Furthermore, it is relevant from a social point of view, as many
curricula provide the opportunity (or, sometimes, requirement) to learn two
languages at school, and yet no firm evidence exists on how this may affect the
outcome of the learning process.

Thus, similarity in form,
when coupled with dissimilarity in meaning, may not always be helpful for
learners.
639 Language Learning 59:3, September 2009, pp. 623–646
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
Specifically, we propose that in the acquisition of L2, when competence
is limited and participants have difficulties in recalling the correct items, two
strategies are employed with reference to the L1: either maximizing the similarities
ormaximizing the differ...

Language Learning 59:3, September 2009, pp. 623–646 623
C 
2009 Language Learning Research Club, University of Michigan
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
Introduction
Several factors affect the learning of a second language (L2).

Language Learning ISSN 0023-8333
L2 Vocabulary Acquisition in Children:
Effects of Learning Method and Cognate
Status
Claudio Tonzar
University of Urbino
Lorella Lotto
University of Padova
Remo Job
University of Trento, Italy
In this study we investigated the effects of two learning methods (picture- or wordmediated
learning) and of word status (cognates vs. noncognates) on the vocabulary
acquisition of two foreign languages: English and German.We examined children from
fourth and eig...

A further observation about the
relative “independence” of two languages learned concurrently involves the rate
of substitutions of the two languages: Very rarely words in one language were
substituted for by words in the other language.

Language Learning ISSN 0023-8333
L2 Vocabulary Acquisition in Children:
Effects of Learning Method and Cognate
Status
Claudio Tonzar
University of Urbino
Lorella Lotto
University of Padova
Remo Job
University of Trento, Italy
In this study we investigated the effects of two learning methods (picture- or wordmediated
learning) and of word status (cognates vs. noncognates) on the vocabulary
acquisition of two foreign languages: English and German.We examined children from
fourth and eig...

Given that the two groups of children
taking part in the study differ in their knowledge of one of the L2s we used
(English), we will not compare directly the results of the two age groups, on
the assumption that it is likely that eighth-grade students would perform better.

Struzzo Ostrich Der Strauß 1.35 2.20 1.55 7 7 6 95 14.60

The shift from one form of representation to the other is assumed
to be gradual and to be composed of changes in the relative weight given to
the lexical and conceptual links: As fluency increases, the existing L2-to-L1
lexical links are complemented, rather than substituted, by stronger additional
links between the L2 lexicon and the conceptual system.

For instance, S´anchez-Casas et al. suggested
that cognates share the same lexical representation, whereas noncognates are
represented as separate entries in lexical memory.

However, in previous studies, more advanced
L2 learners were also older L2 learners; thus, age and proficiency may be
confounded.

Two measures were derived from the two pretests: an index of name agreement
and the mean similarity ratings.We used these indexes to select the experimental
material.

As expected, the worst recall scores were registered in the
635 Language Learning 59:3, September 2009, pp. 623–646
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
last session, 1 month later (English, fourth graders: .45, eighth graders: .62;

Materials and Procedure
The 40 pictures selected on the basis of the two pretests and the corresponding
names in Italian, English, and German were used throughout the study (see the
Appendix).

Furthermore, it is relevant from a social point of view, as many
curricula provide the opportunity (or, sometimes, requirement) to learn two
languages at school, and yet no firm evidence exists on how this may affect the
outcome of the learning process.

The analyses of
variance (ANOVA) on proportion of correct responses were performed after a
2 arcsin transformation.

The issue
of learning two L2s at a time is a theoretically interesting matter because
it touches on the constraints of the acquisition mechanism involved in L2
learning.

Two ways to maximize differences are the following:
3. using letters of L2 that are not part of the L1 alphabet (e.g., inserting “w,”
“y,”or “ß,” which are not part of the Italian alphabet; “Flatoß” instead of
“Fl¨ote”);
4. adapting L1 words to L2 word forms (e.g., deleting a vowel, or adding a
consonant, thus differentiating it from Italian words that generally have a
final-position vowel; “M¨uhler” instead of M¨uhle).

Congruent with Chen’s (1990) data, performance
was better when learning and test condition were congruent, compared to when
they were incongruent.

Congruent with Chen’s (1990) data, performance
was better when learning and test condition were congruent, compared to when
they were incongruent.

The results showed a
significant congruency effect: Participants were more efficient with congruent
stimuli, in other words, the word-learning group was faster in translating in
L2, whereas the picture-learning group was faster in naming pictures in L2.
However, a main effect of learning method was not found.

Analyzing a subset of learners’ errors, we
found that when a cognate was wrongly reported, it conformed more often to
the strategy of maximizing the similarities, wherease when a noncognate was
wrongly reported, it conformed more often to the strategy of maximizing the
differences (see Table 8).

On the one hand,
cognates were found to facilitate the acquisition of L2 words, but when the language was
familiar (i.e.,

Imbuto Funnel Der Trichter 1.20 1.15 1.15 6 6 8 100 0.00

The shift from one form of representation to the other is assumed
to be gradual and to be composed of changes in the relative weight given to
the lexical and conceptual links: As fluency increases, the existing L2-to-L1
lexical links are complemented, rather than substituted, by stronger additional
links between the L2 lexicon and the conceptual system.

Because our aim was mainly to investigate the participants’ quantitative
performance when asked to learn a vocabulary set comprising cognate and
noncognate words under two teaching/learning methods, we will not present a
detailed analysis of learners’ errors.

In spite of the fact
that a between-subjects design makes it more difficult to control for all relevant
learner variables, we preferred it to a within-subjects design for reasons of time
and because it can be compared to many of the studies done in the educational
field.

Objects belonging to 12 semantic categories
(buildings, vehicles, weapons,musical instruments, animals, plants, fish,
kitchen utensils, fruits, tools, clothes and vegetables) were depicted.

In addition, we should reiterate that
such superiority is context-sensitive.

This shows that when noncognates
have been learned to some degree (after the first immediate test session),
the picture-word association is a good cue for recall.

From a theoretical standpoint, this study provides evidence on the differential
role played by the links between the conceptual system and the L1 and L2
Table 8 Proportion of wrongly reported words for cognates and noncognates as a
function of the two strategies
Maximizing similarities Maximizing differences
Cognate Transform L2 into Adapt L2 to Insert peculiar Adapt L1 to
status “canonical” L1 L1 word forms L2 letters L2 word forms
Cognates 27.3 12.5 0.9 3.03

Lexical memory in novice bilinguals: The role of
concepts in retrieving second language words.

The same two learning methods were contrasted by Lotto and De Groot
(1998), who instructed two groups of Dutch undergraduates to learn words in
a novel language (Italian).

Language Learning 59:3, September 2009, pp. 623–646 623
C 
2009 Language Learning Research Club, University of Michigan
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
Introduction
Several factors affect the learning of a second language (L2).

The more often cited study on this issue was
conducted by Potter, So, von Eckhart, and Feldman (1984), who tested two
models of word processing and representation of the corresponding linguistic
knowledge in memory: the word association model and the concept mediation
model.

The Study
Phase 1: Collecting Ratings for Name Agreement and Cognate Status
Prior to the main learning experiment, we performed two pretests.

In a rendition
of an English or a German word, two (possible) ways to maximize similarities
between the two languages are the following:
1. transforming L2 phonemes into “canonical” L1 graphemes (e.g., writing
“Carafe” for “Karaffe,” because in Italian the sound /k/ is spelled as /c/);
2. adapting L2 words to L1 word forms (e.g., adding a vowel in last position:
“Polipe” instead of “Polyp,” because nouns in Italian tend to end with a
vowel).

However, although the concept mediation model also
assumes a direct link between the lexicon of the L2 and the conceptual system,
the word association model does not and it assumes that L2 words gain access
to concepts indirectly through L1 mediation.

Because our aim was mainly to investigate the participants’ quantitative
performance when asked to learn a vocabulary set comprising cognate and
noncognate words under two teaching/learning methods, we will not present a
detailed analysis of learners’ errors.

After a learning phase during which L2
words were presented either together with the corresponding translation in L1 (wordlearning
method) or with the picture of the object to which the word refers (picturelearning
method), children were tested at different points in time.

Pavone Peacock Der Pfau 2.25 1.70 1.45 6 7 4 95 12.16

Several explanations may be offered for these results,
some emphasizing the way in which cognates and noncognates are represented
in memory and some relating to the differences between the two set of words
with regard to retrieval processes.

For half of the participants, English preceded German, and for the
other half, the order was reversed.

Thus, similarity in form,
when coupled with dissimilarity in meaning, may not always be helpful for
learners.
639 Language Learning 59:3, September 2009, pp. 623–646
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
Specifically, we propose that in the acquisition of L2, when competence
is limited and participants have difficulties in recalling the correct items, two
strategies are employed with reference to the L1: either maximizing the similarities
ormaximizing the differ...

Several explanations may be offered for these results,
some emphasizing the way in which cognates and noncognates are represented
in memory and some relating to the differences between the two set of words
with regard to retrieval processes.

In our study, the cognate stimuli were cognates in all three languages, and this
would maximize the possibility of detecting a difference between cognates and
noncognate stimuli.

For noncognates, this covert translation
would not result in a facilitating effect.

Language Learning 59:3, September 2009, pp. 623–646 624
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
The results of the few studies that have focused on the role of these two
different methods of vocabulary acquisition challenge current models of lexical
representation in bilinguals.

The first
consisted of a picture-naming-in-L1 task and was used to make sure that the
pictures we selected were consistently named with a single name.

In the cognate-rating study, we presented 63 Italian-English, Italian-
German, and English-German word pairs corresponding to the names of the
pictures used in the first pretest and asked 20 third-year middle school Italian
children (13 years old) to evaluate the orthographic similarity within each word
pair on a 7-point scale.
627 Language Learning 59:3, September 2009, pp. 623–646
Tonzar, Lotto, and Job L2 Vocabulary Acquisition in Children
Table 1 Mean cognate ratings and mean leng...

Furthermore, it is relevant from a social point of view, as many
curricula provide the opportunity (or, sometimes, requirement) to learn two
languages at school, and yet no firm evidence exists on how this may affect the
outcome of the learning process.

Chen (1990) compared two groups of Cantonese-speaking college
students who were instructed to learn words in a novel language (French) using
two different learning methods.

The first
consisted of a picture-naming-in-L1 task and was used to make sure that the
pictures we selected were consistently named with a single name.

As a consequence, it may
be plausible to hypothesize that in learning a new language, setting up a new
entry in memory is more demanding than modifying a preexisting similar
entry, and this would facilitate the learning of L2 cognates over noncognates.

Chen (1990) compared two groups of Cantonese-speaking college
students who were instructed to learn words in a novel language (French) using
two different learning methods.

Each child participated in four English sessions and four German
sessions.

From an educational perspective,
the results suggest that reliance on, and strengthening of, the link between
the concepts presented as pictures or real objects and their names may help in
vocabulary learning.

This pattern of results appeared to be very robust (for reviews, see Chen,
1992; De Groot, 1993; Kroll, 1993; Kroll & De Groot, 1997; Kroll, Michael,
& Sankaranarayanan, 1998; Kroll & Tokowicz, 2001) and led Kroll and Stewart
(1994) to propose the revised hierarchical model (RHM), which assumes
that less fluent bilinguals process words in the L2 by relying on their lexical
knowledge in the L1, whereas more proficient bilinguals use the concept-to-L2
link to process words without activ...

On the one hand,
cognates were found to facilitate the acquisition of L2 words, but when the language was
familiar (i.e.,

Mean .29 .43 .27 .25 .31
and performance levels on the test following the first learning session and on the
test following the 1-week session were intermediate (.34 and .32, respectively).

The second
pretest was designed to assess the cognate relationship between Italian words
and their translations, both in English and in German.

Students in neither grade
had been systematically exposed to German, although eighth graders had been
taught English in the previous 2 years, albeit on a very restricted schedule.

Thus, on this interpretation, the
interlexical relationship underlies the facilitating effect of the verbal method
for cognates.

The shift from one form of representation to the other is assumed
to be gradual and to be composed of changes in the relative weight given to
the lexical and conceptual links: As fluency increases, the existing L2-to-L1
lexical links are complemented, rather than substituted, by stronger additional
links between the L2 lexicon and the conceptual system.

The picture-naming pretest involved 142 pictures, some of which were
different versions of the same object.

The analyses of
variance (ANOVA) on proportion of correct responses were performed after a
2 arcsin transformation.

The automatic cognate form assumption: Evidence for the parasitic
model of vocabulary development.

The issue
of learning two L2s at a time is a theoretically interesting matter because
it touches on the constraints of the acquisition mechanism involved in L2
learning.

Up to now, most of the studies cited in the
debate focus on the acquisition of L2 from birth, although empirical evidence
on how learning two languages may affect the overall outcome is scanty.

Up to now, most of the studies cited in the
debate focus on the acquisition of L2 from birth, although empirical evidence
on how learning two languages may affect the overall outcome is scanty.

Results
In Figure 1, the percentages of correct responses are plotted as a function of
learning method, cognate status, language, and grade.

The shift from one form of representation to the other is assumed
to be gradual and to be composed of changes in the relative weight given to
the lexical and conceptual links: As fluency increases, the existing L2-to-L1
lexical links are complemented, rather than substituted, by stronger additional
links between the L2 lexicon and the conceptual system.

In this article, we present a study aimed
at investigating how some of these factors affected vocabulary acquisition in a
school environment in which two foreign languages are acquired in close temporal
contiguity.

Internet: claudio.tonzar@
uniurb.it

Two measures were derived from the two pretests: an index of name agreement
and the mean similarity ratings.We used these indexes to select the experimental
material.

Our aim was to try to
single out the effect of learning method in relation to the linguistic distinction
between cognate and noncognate words and to the real-world situation in which
two L2s are learned at the same time.

The shift from one form of representation to the other is assumed
to be gradual and to be composed of changes in the relative weight given to
the lexical and conceptual links: As fluency increases, the existing L2-to-L1
lexical links are complemented, rather than substituted, by stronger additional
links between the L2 lexicon and the conceptual system.

This pattern
of a stronger superiority of the picture-learning method in the latter sessions
shows that the effects of the picture-mediated learning method may be fully
appreciated when the learning sessions are spaced in time.

First,
let us rule out a possible strategic base.

Students in neither grade
had been systematically exposed to German, although eighth graders had been
taught English in the previous 2 years, albeit on a very restricted schedule.

As can be seen, the learning
method showed a consistent effect, with the picture-learning method being
more effective than the word-learning method.

For both methods, the learning phase included two learning sessions for
English and two learning sessions for German held 1 week apart.

In regard to the learning method, in the
educational domain the between-subjects design is quite common for studies of
this kind.

From an educational standpoint, this suggests that more time and effort should
be devoted to noncognates in vocabulary acquisition, even though cognates
may themselves present (different) problems (see below).

Students in neither grade
had been systematically exposed to German, although eighth graders had been
taught English in the previous 2 years, albeit on a very restricted schedule.

In particular, a
marked decrease was observed for noncognates in the delayed tests.