(4) a.
The detective hoped that the lieutenant knew [which accomplice [the
shrewd witness would recognize in the lineup]].
b. The lieutenant knew [which accomplice [the detective hoped that [the
shrewd witness would recognize in the lineup]]].
Complexity rating for those sentences indicate that native speakers of English
judge (4b) as more difficult than (4a), showing that it is more costly for readers to
process sentences with long dependencies. If we use a memory-based account, this
locality effect observed at the verb shows that it is more costly to reactivate the
element stored in working memory at the point of integration, if the activation
level of the element in working memory has decayed (Lewis, 1996, Le wis, Vasishth ,
and Van Dyke, 2006, Van Dyke and Mc Elree, 2006, Wagers, 2008).
4 Expectation and locality: their connections
In previous studies, facilitations of processing caused by expectations are mostly
observed in head-final languages, such as German, Hindi, or Japanese. The
increase of the processing cost caused by locality is often reported for head-initial
languages like English. However, there are some recent suggestions that it is not
quite accurate to consider that locality effects are totally absent from head-final
languages, or expectations do not play a substantial role in head-initial languages.
We review some recent studies of head-final languages that obser ved expectations
and locality effects in the same types of constructions in the same language.
Levy and Keller (2013), critically reviewing Konieczny (2000 ) and Konieczny
and D öring (2003), investigated the verb-final structures in German. They used
materials in which the position of the critical verb, as well as the plausibility of
the sentences, was controlled in their eye-tracking experiments. The templates
of their materials are shown below. They created four versions of target sentences
using the same phrases, by putting an adverbial phrase and a dative NP either
in the temporal subordinate clause (with nachdem ‘after’ in the example below),
or in the relative clause that modifies the subject in the matrix clause (der
Mitschüler ‘the classmate”). The adverbial phrase and the dative NP used in the
target sentences were not a simple phrase but contained several words; thus,
the placement of those phrases altered the length of the clause a lot. According
to Levy and Keller (2 013), verbs in the target sentences (both in the temporal
clause and the relative clause) were all optionally ditransitives; the dative NP
can appear with the verb as an argument, but it may be absent.
(5) The template for the target sentences
a. adverbial and dative NP in subordinate clause
After NP-NOM [Adv] [NP-DAT] NP-ACC verb, NP-NOM [who NP-ACC
verb],...
b.dative NP in subordinate clause; adverbial in relative clause
After NP-NOM [ NP-DAT] NP-ACC verb, NP-NOM [who [Adv] NP-ACC
verb],...Make a good prediction 87