Íslenskt mál og almenn málfræði - 01.01.2023, Page 200

the highest rate. If the equivalence was clear, there would be very little line over-
lap in the graph.
Instead, systematic contrast replication at the group and individual level is
observed, or regularization within matched conditions (Figures 6 and 7). In other
words, children acquire the conditioning patterns observed in their caregivers’
results without producing the variants at the same rates.
The children’s rate of DS is always higher, pointing towards incrementation.
This is in line with results from previous work (e.g. Smith et al. 2007, Hall and
Maddeaux 2020 and Repetti-Ludlow and MacKenzie 2022) and has implications
for studies on the role of variant specialization (Wallenberg 2019) and the
dynamics of variation in individuals (Tamminga et al. 2016) for the directionality
of change. I furthermore argue that models of language acquisition need to be
able to account for the emergence of productive intra-speaker variation patterns.
In this context, I show that a combination of Yang’s Variational Model of Language
Acquisition (2002), where children can acquire competing rules, and the Tolerance
Principle (2016) is a promising approach in that respect (as in Woods et al. 2021),
with grammatical conditioning of variation being reframed in the context of pro-
ductive rule discovery. Additionally, I argue that a case should be made for sto-
chastic exceptions in addition to stochastic rules.
Iris Edda Nowenstein200
Figure 6: Dative selection/acceptance in the first and third person singular, by
group (children and caregivers).
Figure 7: Dative selection/accusative rejection with/without syncretism on the
subject, by group (children and caregivers).
R
at
e
of
a
cc
us
at
iv
e
re
je
ct
io
n:
P
ar
en
ts
R
at
e
of
d
at
iv
e
ac
ce
pt
an
ce
: P
ar
en
ts
R
at
e
of
d
at
iv
e
se
le
ct
io
n:
C
hi
ld
re
n
R
at
e
of
d
at
iv
e
se
le
ct
io
n:
C
hi
ld
re
n
no syncretism syncretism
1st.pers.sing 3st.pers.sing 1st.pers.sing 3st.pers.sing
no syncretism syncretism