Processing Morphology in reading Hungarian

1
Processing Morphology in reading Hungarian

• Csaba Pléh and Ágnes Lukács
• Collegium Budapest and
• Center for Cognitive Science Budapest U. of
  Technology and Economics

2
Some reading relevant aspects of agglutination

• Processing
• Segmentation
• Lexical access
• holistic, analytic
• Allomorphy

• Reading aspect
• Long words
• Analytic effects
• reading vs. acoust
• Rules in reading

3
Studies and issues

• Lexical decisions critical points
• Multiple activation crossmodal
  semantic spread
• Allomorphy rules and modality

4
Lexical decisions(Juhász and Pléh)

• UNFOLDING OF A WORD
• ki
• ki- véd-
• ki- véd- és -
• ki- véd- és é-
• ki- véd- és é - re
• 'on its defense'

5
Different spoiling

• fel-védésére 'on its upfense' prefix wrong
  combination
• ku-védésére prefix non existent
• kivédésé-ra case non harmonizing
• kivédésé-ge case non existent
• Non existing is much faster to reject

6
General implications

• Decision times were extremely slow.
• 450 ms by syllables compared to 2-300 by Taft
  (1994) and Laudanna et al. (1994).
• Representation rather than access effects
  (Marslen-Wilson et al., 1994)
• Or slower reading process due to late
  grammatical informations over words.
• Stronger reliance on analytic strategies in an
  agglutinative language.  

7
Overall results
 
8
Activation of dual meanings in reading (Gergely,
Thuma, and Pléh)

• DOB1 drum DOB2 throw
• Stems, ambiguous suffixes and dismbigauting
  suffixes
• Biased instrument U toss
• Stem Az ütosök közt János kedvence a DOB.
• Among the percussionists John's favorite is the
  DRUM'
• Ambig. suffix Utóljára János kezében láttam a
  DOBOD.
• 'Last time I saw your DRUMyou throw in John's
  hand
• Disambig. suffix Pista egy hatalmasat ütött a
  DOBRA.
• 'Steve had a big hit ON THE DRUM' 

9
Crosslanguage facilitation

• Azt hittem, valaki csönget.
• I thought someone rings
• relevant phone
• irrelevantgold

10
Crossmodal facilitation across languages
11
Crossmodal facilitation and the rule debates
(Lukács and Pléh)

• Many allomorphy subtypes in Hungarian
• That provides a good site for testing the
  implications of the rule issue
• At the same time this involves modality neutral
  representation

12
Priming effects and alomoprphy types
13
What all this implies

• All questions, non answers
• Is agglutination slowing reading?
• Crossing from L1 sound to L2 writtten?
• The modality issue in rule rekated priming is non
  trivial