Transcript Radboud University Nijmegen - Low Educated Second Language

The interpretation of inflectional suffixes by loweducated L2 Dutch learners
SLRF, Pittsburgh, October 2012
Loes Oldenkamp, Radboud University Nijmegen, the Netherlands
Background
L2 forms (3sg):
- kom – t
- *kom – Ø
- *kom – ə(n)
- *kom – ət
- *kom – tə
vader *komet
met
father come-3sg with
zijn dochter *fietset
his daughter cycle-3sg
Low-Educated Second Language & Literacy Acquisition
(LESLLA)
• Learners have only received (some years of) primary education in their home
country.
• Often non-literate in the Latin script (Moroccan and Chinese learners).
• Little support of written language for most learners.
• No knowledge of other language than their L1.
• They have no (or few) metalinguistic skills.
• Language acquisition proceeds (often) slowly.
Background
•L1 acquisition
Inflectional morphology is acquired smoothly.
•L2 acquisition
Acquisition of inflectional morphology appears to be a persistent problem.
Dutch inflectional morphemes
V-Verb
N-Noun
Ø
t
ə(n)
ə(n)
s
Dutch inflectional morphemes
Properties:
1. They consist of a coronal (/t/, /s/) or a schwa
2. They are invisible for stress
 Adding these morphemes lead to complex word forms:
 Adding a schwa to the stem of a word leads to polysyllabic words
e.g. loop + schwa = lope(n)
 adding a coronal to the stem leads to a final consonant cluster
e.g. loop + /t/ = loopt
Avoidance strategies
The L2 learner avoids words ending in a /t/-final consonant cluster and
polysyllabic words ending in schwa:
/t/-final consonant cluster:
polysyllabic words ending in schwa:
-/t/ deletion at the end of a word
- deletion of schwa
e.g. loop instead of loopt
- schwa insertion
e.g. loopet or loopte instead of loopt
e.g. loop instead of loope(n)
Possible sources
•L1 morphosyntax
•L1 phonology
Similarities and dissimilarities between Dutch, Turkish (TU),
Moroccan Arabic (MA) and Mandarin Chinese (CHIN)
Dutch
TU
MA
CHIN
+
+
+/+/-
-
-
SVO/ SOV
+/+
SOV
+
+/-
SVO/ VSO
+
+
SVO
-
•phonology
- consonant cluster in coda
- word-final schwa
•morphosyntax
- word order
- verbal morphology
- nominal morphology (plural)
L2 learners have difficulties producing inflectional morphology correctly,
but:
Do L2 learners have difficulties in interpreting inflectional
morphology correctly as well?
Method (participants)
•
Picture selection task: participants are orally provided with a stimulus
and have to select the corresponding picture.
•
44 Turkish, 44 Moroccan Arabic and 42 Mandarin Chinese learners of
Dutch participated in the experiment.
•
No more than three years of secondary education in their home
country.
•
Level A1 (Basic User: Breakthrough), A2 (Basic User: Waystage) and
B1 (Independent User: Threshold), (CEF).
Method (materials)
•
110 items: 54 target items and 56 distracter items.
•
Target items: nouns (singular vs. plural) and verbs (3rd person singular
vs. 3rd person plural).
•
The only cue to interpret the utterance correctly is the inflectional ending.
Ze kust een jongen.
she kiss-3SG.PRES a boy
Ze kussen een jongen.
she kiss-3PL.PRES a boy
-
de kip
de kippen
the chicken-SG
the chicken-PL
Categories
1. lexical items, nouns
e.g. de pop/ de kip
the doll-SING/ the chicken-SING
2.
lexical items, verbs
e.g. ze koopt een boek./ ze kust een jongen.
she buy-3SG.PRES a book/ she kiss- 3SG.PRES a boy

3.
Lexical items: constructed to test whether participants knew the vocabulary of
the nouns and verbs that were used in the experiment.
inflectional items, nouns
e.g. de kat/ de katten
the cat- SING/ the cat-PLUR
4.
inflectional items, verbs
e.g. ze wast een auto./ ze wassen een auto.
she wash-3SG.PRES a car/ she wash-3PLUR.PRES a car

Inflectional items: constructed to test nominal inflection (singular vs. plural) and
verbal inflection (3rd person singular vs. 3rd person plural, present tense).
Lexical item, noun
Stimulus: de pop
the doll-SING
Lexical item, verb
Stimulus: Ze kust een jongen.
She kiss-3SG.PRES a boy
‘She kisses a boy.’
boy
Inflectional item, noun
Stimulus: de katten
the cat-PLUR
Inflectional item, verb
Stimulus: Ze wast
een auto.
she wash-3SG PRES a carr
‘She washes a car.’
Results (overall)
Lexical (k=24)
Inflection (k=25)
Mean
Mean
Example of item pairs
noun (k=14)
.968
de pop / de kat (‘the doll / the cat’)
verb (k=10)
.916
ze loopt / ze lacht (‘she walks / she laughs’ )
total (k=24)
.942
noun (k=14)
.723
de pop / de poppen (‘the doll / the dolls’)
verb (k=11)
.326
ze loopt / ze lopen (‘she walks / they walk’)
total (k=25)
.524
noun (k=28)
.845
verb (k=21)
.621
Mean (k=49)
.733
1 = lexical item pairs
2= inflectional item pairs
Results (per category)
• Separate analyses on the 4 different item types
lexical items, nouns;
lexical items, verbs;
inflectional items, nouns;
inflectional items, verbs.
• Univariate ANOVAs with ‘proportions accurate responses’ as
dependent variable and ‘L1 background’ (Turkish, Moroccan
Arabic or Mandarin Chinese) and ‘L2 proficiency level’ (A1, A2 or
B1) as independent, fixed factors.
Lexical items, nouns
A1
A2
B1
Mean
Turkish
.943
.944
.990
.959
Moroccan
.962
.964
.976
.968
Chinese
.975
.981
.976
.977
Mean
.959
.964
.981
.968
• Participants performed very well on these items.
• No significant effects for ‘L1 background’ (F (2, 119) = 1.14, n.s.) and ‘L2
proficiency’ (F (2, 119) = 1.72, n.s.). and did not differ from each other with
respect to ‘L1 background’ and ‘L2 proficiency’.
Lexical items, verbs
A1
A2
B1
Mean
.913
.943
.979
.944
Moroccan .907
.936
.933
.925
Chinese
.814
.900
.925
.878
Mean
.880
.926
.946
.916
Turkish
• Participants performed very well on these item types.
• Significant main effects for both ‘L1 background’ (F (2, 119) = 6.07, p < .01,
η2 = .093) and ‘L2 proficiency’ (F (2, 119) = 6.76, p < .01, η2 = .102).
• No significant interaction (F (4, 19) = .994, n.s.)
Inflectional items, nouns
A1
A2
B1
Mean
.662
.745
.924
.774
Moroccan .691
.832
.852
.791
Chinese
.531
.595
.673
.596
Mean
.630
.721
.824
.723
Turkish
• Significant effect for ‘L1 background’ (F (2, 119) = 8.69, p< .000, η2 = .127):
• Chinese participants performed worse than Turkish and Moroccan
participants.
• Turkish and Moroccan participants did not differ from each other.
• Significant effect for ‘L2 proficiency’ (F (2, 119) = 6.87, p= .000, η2 = .104):
• Participants with level B1 performed better than participants with level A1,
but not better than participants with level A2.
• A1 and A2 did not differ.
• No interaction
Inflectional items, verbs
A1
A2
B1
Mean
.220
.286
.501
.336
Moroccan .161
.297
.565
.341
Chinese
.211
.295
.397
.301
Mean
.197
.293
.488
.326
Turkish
•
Significant effect for ‘L2 proficiency’ (F (2, 119) = 16.47, p< .000, η2 = .217).
•
No effect for ‘L1 background’: all participant groups appeared to have
difficulties interpreting these items.
Conclusions (1)
• Participants have more difficulties in the interpretation of
inflectional items than in the interpretation of lexical items.

Redundant information is often available in the input. Learners are not
used to focusing on the inflectional ending and having the inflectional
ending as the only cue in interpreting the stimulus correctly.
e.g.
a. de stoel
the chair
b. de stoel-en
the chair-PLUR
=> No redundant information available.
a. één stoel
one chair
b. drie stoel-en
three chair-PLUR
=>The numerals één and drie make the plural morpheme redundant.
Conclusions (2)
• Participants have more problems interpreting verbal inflection (=
contextual inflection) correctly than in interpreting nominal inflection
(= inherent inflection).

Nouns and verbs differ in semantic complexity: The nominal plural
morpheme only expresses number; the 3rd person singular, present tense
and 3rd person plural, present tense morphemes express person, number
and tense.
Conclusions (3)
• L2 proficiency level clearly plays a role: The interpretation of
morphological elements improves with proficiency, in all groups of
learners.
Conclusions (4)
• L1 background plays a role.

Mandarin Chinese participants had more difficulties interpreting
inflectional endings than Turkish and Moroccan Arabic participants, but
not in verbs. All participant groups appeared to have difficulties in verbal
inflection.
Loes Oldenkamp
Radboud University Nijmegen, the Netherlands
[email protected]