Written morphological analogies in Hebrew language development

Document Sample
Written morphological analogies in Hebrew language development Powered By Docstoc
					Written morphological analogies
          in Hebrew



       Dorit Ravid and Rachel Schiff
   Tel Aviv University Bar Ilan University
                    Israel
                  Morphology

• One of the organizing principles of the mental
  lexicon
• Morphemes and words
                   Morphology

• One of the organizing principles of the mental
  lexicon
• Morphemes and words
• Crucial importance in Hebrew
   – Highly synthetic Semitic language
               Hebrew morphology:
                Roots and patterns
katav hixtiv katuv mixtav kituv
wrote    dictated written    letter    subtitle


któvet ktav ktiv katvan katava
address writing spelling typist journalistic report


hitkatvut            ktuba             taxtiv
correspondence marriage contract      dictate
    Roots and patterns: k-t-b ‘write' ‫כתב‬
katav hixtiv         katuv mixtav kituv
wrote    dictated    written      letter      subtitle
‫כתב‬      ‫הכתיב‬       ‫כתוב‬        ‫מכתב‬          ‫כיתוב‬
któvet ktav ktiv katvan katava
address writing spelling       typist      journalistic report
‫כתובת‬     ‫כתב‬       ‫כתיב‬        ‫כתבה כתבן‬
hitkatvut           ktuba                  taxtiv
correspondence marriage contract           dictate
‫התכתבות‬             ‫כתובה‬               ‫תכתיב‬
         Roots (k-t-b, g-d-l, p-r-s-m)
     Spoken roots                    Written roots
• Discontinuous               •   Almost continuous
• 3-4 consonants              •   3-4 letters
• Phonological alternations   •   Consistent orthography
• Lexical core of             •   Construal as entity
  morphological family            fostered by written
• Salient                         properties
      Patterns (hiCCiC, CaCuC, miCCaC)
       Spoken patterns               Written patterns
•   Discontinuous             •   Scant orthographic
•   Prosodic templates            representation
•   Provide internal vowels   •   Vowels: almost no
    (+ prefixes / suffixes)       representation
•   Categorial meaning        •   Discrete prefix, suffix
    – Verbal                  •   Construal as entity
    – Nominal                     obscured by written
• Less salient than roots         properties
  Implications for reading and writing
• The lexically meaningful part of the word is
  represented in its center; letters framing the
  word carry grammatical and categorial meaning
• WKŠBMGDLYKM ‫וכשבמגדליכם‬
  u-xshe-be-migdaley-xem ‘and-when-in-towers-
  yoursPl’
• Root GDL ‘grow’ surrounded by function
  elements
              The current study
• Investigates Hebrew readers’ ability to analyze
  roots and patterns in written Hebrew wordforms
• Focus on nominal patterns
• Testing the ability to extract and recombine roots
  and patterns from written Hebrew nominals
  using a morphological analogies task
                      Participants
1. 152 gradeschool children, middle-high SES
2. 167 gradeschoo children, low SES
   –   Five age-groups each: 2nd, 3rd, 4th, 5th and 6th graders
   –   All monolingual speakers of Hebrew as a native tongue with
       no diagnosed language or learning disabilities
3. 38 undergraduate education students, all with a long
   history of reading difficulties and diagnosed with
   reading disabilities within three years prior to attending
   university or while attending university; All monolingual
   speakers of Hebrew as a native tongue
    The Morphological Analogies Task
                (MAT)
• 40 analogy sets
• Each set contains two components:
   – A set of stimulus nouns
   – A set of possible responses
• The task requires the selection of a target noun
  from the set of responses to complete the
  stimulus set
                Analogy structure
P                      Root relationship                      P
a                                                             a
t
t
         KPL ‫כפל‬                       MKPLH ‫מכפלה‬            t
                                                              t
e
r
    kéfel ‘multiplication’            maxpela ‘multiple,N’    e
                                                              r
n                                                             n

r
e                                       Pattern source        r
                                                              e
l
a       SRŦ ‫סרט‬                                ?              l
                                                              a
t
        séret ‘film’
                                                              t
i                                      (MSRŦH ‫)מסרטה‬          i
o                                                             o
n                                     (masreta ‘projector’)   n
s                                                             s
h       Root source                                           h
i                                                             i
p                      Root relationship                      p
                       Response set
1.   Correct response: target noun MSRŦH ‫‘ מסרטה‬projector’
2.   Main root distracter: a word containing the same root as the root
     source, but not the same pattern TSRYŦ ‫‘ תסריט‬script’
3.   Pattern distracter: a word containing the same pattern as the
     pattern source, but not the same root MGRPH ‫‘ מגרפה‬rake’
4.   Secondary root distracter: a word containing the same root
     shared by members of the top horizontal pair KPYL ‫‘ כפיל‬double’
5.   Semantic distracter: associated semantically or pragmatically but
     not morphologically to left-hand member of horizontal pair
     KWLNW9 ‫‘ קולנוע‬movies’
                                 Correct responses:
                      100
                            High and Low SES gradeschoolers
                       90

                       80                H-SES
% correct responses




                       70

                       60

                       50

                       40
                                         L-SES
                       30

                       20
                                                      5/6 > 3/4 > 2
                       10
                                                      High > Low
                        0

                            G2      G3           G4     G5        G6
                                           Correct responses:
                      100
                            High and Low SES gradeschoolers, adult dyslexic students
                       90

                       80              H-SES
% correct responses




                       70

                       60
                                                                     Ad dys
                       50

                       40                 L-SES
                       30

                       20

                       10

                        0

                             G2      G3        G4      G5       G6      Adult
                                                                      dyslexics
                                     Correct responses:
                      High and Low SES gradeschoolers, adult dyslexic students
                      100

                       90

                       80                 H-SES
% correct responses




                       70

                       60
                                                                     Ad dys
                       50
                                             L-SES
                       40

                       30

                       20

                       10

                        0

                            G2       G3       G4       G5       G6      Adult
                                                                      dyslexics
       Erroneous response types: High SES

100%

80%
                                 Semantic distracter
60%                              Secondary root distracter
                                 Pattern distracter
40%                              Main root distracter


20%

 0%                              Main root response
                                 Few pattern, semantic responses
       G2    G3   G4   G5   G6   No age differences
       Erroneous response types: Low SES
       Main root: increase with age; Secondary root: decline with age
       Semantic and pattern distracters: decline with age
100%
 90%
 80%
 70%                                         Semantic distracter
 60%                                         Secondary root distracter
 50%                                         Pattern distracter
 40%                                         Main root distracter
 30%
 20%
 10%
  0%
       G2      G3      G4      G5      G6
Erroneous response types: Adult dyslexic students

100%
 90%
 80%
                                 Semantic distracter
 70%
                                 Secondary root distracter
 60%
                                 Pattern distracter
 50%
                                 Main root distracter
 40%
 30%
 20%                          Main root distracter
                              Semantic distracter
 10%
  0%
        Summary and conclusions
• Analytical morphological skills from early on in
  normally developing Hebrew speakers
   – More in high-SES gradeschoolers, less in low-SES
• Dyslexics are ‘stuck’ with the analytic skills of
  3rd and 4th graders
   – Revert to non-morphological strategies absent in
     typically-developing children
        Summary and conclusions
• Roots perceived as the prime lexical construct in
  Hebrew words
• Patterns less salient and their perception lags
  behind that of roots
• But - impossible to solve the MAT without
  recourse to both root and pattern
                                        Correct responses:
                            High SES gradeschoolers, real versus nonce words
                      100


                      90


                      80


                      70
% correct responses




                      60


                      50


                      40


                      30


                      20    Nonce: 5/6 > 2/3/4                      Real words
                      10    Real > nonce                            Nonce words

                       0

                               G2        G3         G4         G5          G6
Erroneous response types: High-SES, nonce words
100%
 90%
 80%
 70%
 60%                               Secondary root distracter

 50%                               Pattern distracter
 40%                               Main root distracter
 30%
 20%
 10%
  0%                             Main root: decline with age
                                 Pattern: surge in 5/6
       G2   G3   G4    G5   G6

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:7
posted:7/4/2012
language:English
pages:23