19 International Symposium on Theoretical and Applied Linguistics ...

G E N E R A L S E S S I O N
L1 transfer in L2 learning: compound forms in the speech of Turkish learners of Greek
Marina Tzakosta
University of Crete/ Greek State Scholarships foundation
martzak@gmail.com
Greek compounding has been extensively dealt with theoretically (cf. Ράλλη <strong>19</strong>99, Nespor & Ralli
<strong>19</strong>96, Ralli 2005); however, little has been said regarding compound formation in second language
learning (L2). In this study, we evaluate the capacity of Greek native speakers and Turkish learners of
Greek to form compound words placing emphasis on a) their internal structure, b) their headedness, c)
the relation holding between the compound elements and d) the status of the linking vowel. In an offline
task, 40 native speakers of Greek (age range: 18-58 years), had to form a) existing compounds
(test 1) and b) morphologically possible but semantically vague compounds (test 2). The same test
was distributed to Turkish learners of Greek (age range: 10-12 years).
Native speakers exhibit a high degree of variation in the formation of endocentric compounds (1a).
Therefore, the syllable receiving stress and the inflectional suffix vary. In case the second compound
constituent retains its inflectional suffix, a syllable of the second compound member receives stress. If
the second compound member does not retain its inflectional suffix, the linking vowel is stressed.
Variation is also attested in compounds of different classes, i.e. [[stem+stem]+deriv.Suff.] and
[stem+word] forms (cf.Nespor&Ralli <strong>19</strong>96, Ralli 2005), which are used interchangeably (1b).
Apparently, the two compound classes are not recognized as distinct. Additionally, the head of the
word is often ambiguously perceived. In (1c) the actual head appears either at the left or right edge of
the word, depending on whether the head has been perceived as the head of the word or not (cf.
Jarema et al. <strong>19</strong>99, Kehayia et al. <strong>19</strong>99, Tsapkini et al. <strong>19</strong>99). Head misperception leads to semantic
ambiguity. Finally, the linking vowel appears in environments where it could be phonologically
prohibited (second example in (1d)). Variable forms appear more systematically in test 2 compared to
test 1. Obviously, variation in test 1 is restricted by mnemonic mechanisms which drive compound
formation.
L2 data display equivalent patterns of variation. However, compounding in L2 is clearly influenced
by L1 word formation mechanisms. Therefore, compound heads are ambiguously perceived not only
in exocentric but also in endocentric compounds (2). A related outcome of Greek L2 is that Turkish
learners prefer two-word compounds to one-word compounds. Additionally, given that linking vowels
do not appear in Turkish compounds, Turkish learners exhibit variable forms with and without linking
vowels (examples in 2a, 2c). Moreover, stress may shift beyond the trisyllabic window, given that
trisyllabicity is not respected in Turkish (first example in 2c). Statistically, [[stem+stem]+deriv.Suff]
forms are more prevalent in L2. This prevalence is theoretically founded in that Turkish compounds
lack inflectional suffixes.
In sum, the degree of variation in compound formation is determined by the extent to which
mnemonic mechanisms are activated. Mnemonic knowledge, which prohibits variation in the formation
of existing words, is minimized in the formation of non-existing words. Variation in L2 compound
learning is additionally driven by mechanisms active in the learners’ L1.
(1) a. ασχηµ-ο-πούλ+ι vs. ασχηµ-ό-πουλ+ο ‘ugly bird–NEUT.NOM.SG.’(Test 1)
b. βαρι-ό-σκαλ+α vs. βαρι-ο-σκάλ+α ‘heavy staircase–FEM.NOM .SG.’(Test 2)
c. κρεατ-ό-ρυζο vs. ρυζ-ό-κρεας ‘meat rice–NEUT.NOM.SG.’ (Test 1)
head head
d. νερ-Ø-έµπορος vs. νερ-ο-έµπορος ‘water seller–MASC.NOM.SG.’(Test 2)
(2) a. βαρι-Ø-σκάλα vs. βαρι-ο-σκάλα vs. σκαλ-ό-βαρο ‘heavy staircase–FEM.NOM.SG.’ (Test 2)
b. κρε-ό-ρυζ+ο vs. κρε-ο-ρύζ+ο/-ι ‘meat rice–NEUT.NOM.SG.’ (Test 1)
c. νερ-ό-εµπoρος vs. νερ-Ø-έµπορος vs. εµπορ-ό-νερος ‘water seller–MASC.NOM.SG.’(Test 2)
Exploring the influence of the interview
as a research method on the construction of identities in narrative
Dorien Van De Mieroop
Lessius / KULeuven
dorien.vandemieroop@lessius.eu
Narratives are an important means by which individuals make sense of their lives and by which they
perceive and construct identities. On the basis of social constructivist insights, this self is regarded as
<strong>19</strong> th ISTAL 59