Native English speakers learning Arabic:
The influence of novel orthographic informa tion
on second language phonological acquisition
Catherine E. Showalter
Indiana University, Department of Second Language Studies
EUROSLA, University of Amsterdam
August 2013
Introduction
• Orthographic input can hinder phonological acquisition (e.g., Bassetti,
2006; Hayes‐Harb, Nicol, & Barker, 2010)
• Orthographic input can aid phonological acquisition (e.g., Escudero,
Hayes‐Harb, & Mitterer, 2008)
• Unfamiliar orthographic symbols can also aid learners (Showalter &
Hayes‐Harb, 2013)
Showalter & Hayes‐Harb (2013)
• Previous studies utilized familiar orthographic symbols and relied on
learners’ knowledge of existing L1 grapheme‐phoneme correspondences
• Can native English speakers use orthographic tone marks to remember
the lexical tones associated with new L2 words?
• L1 English, L2 pseudo‐Mandarin
• Mandarin four‐way lexical tone contrast (tones 1,2,3,4)
• Pinyin: Romanized Mandarin writing with diacritic tone marks, e.g. <fiàn>
Experiment 1
Word Learning Phase Example: Hear [ɡi‐tone1], See:
No Tone Marks Tone Marks
Test Phase Example: Hear [ɡi‐tone2], See:
Respond: Matched or Mismatched
Showalter & Hayes‐Harb, 2013
Experiment 2
• Did subjects learn what tones are indicated by each tone mark or did
they simply pay better attention to the auditory forms because they
notice that tone marks differed across words?
• Stimuli: all stimuli used were the identical to Experiment 1
• Procedures: all procedures from Experiment 1 were the same except
the Final Test.
• Orthographic representations NOT nonobject pictures
• Same number of mismatched/matched items
Mismatched: [fian]‐ tone 3
Showalter & Hayes‐Harb, 2013
Showalter & Hayes‐Harb Findings
• On items that probed sensitivity to tone contrasts, subjects who saw
tone marks significantly outperformed those who did not
• The availability of orthographic tone marks can help learners
remember lexical tone
• Learners can create novel tone‐tone mark correspondences to some extent
• Even unfamiliar orthographic symbols can influence L2 word form
learning
Showalter & Hayes‐Harb, 2013
Present Study
• We do not yet know whether learners can use an entirely unfamiliar
orthography to make inferences about the phonological structure of
L2 words…
• Can learners use their knowledge that letters represent sounds to
infer phonological contrasts in new L2 words even when the
orthography is entirely unfamiliar?
Showalter MA Thesis, 2012
>ططططControl Condition: meaningless Arabic sequence <•
Experiment 1
• L1 English, L2 pseudo‐Arabic
• Two conditions (n=15 in each)
• Orthography Condition: Arabic script
• Materials
• Arabic velar‐uvular contrast /k‐q/
• Words produced by two male native Jordanian Arabic speakers
• 6 nonword minimal pairs, e.g., [kubu]‐[qubu]
Showalter MA Thesis, 2012
Word Learning Task
• Each of the nonwords was associated with a different picture
• E.g., for the auditory forms [kubu] and [qubu], subjects in the Orthography
group saw:
• And the control group saw:
Showalter MA Thesis, 2012
Word Learning Task
• Each auditory word and corresponding picture presented twice per
block (24 items per block)
• Block presented four times, in a different random order each time
and for each subject
• No response required of subjects; told to learn the words and their
meanings as well as possible
• Word Learning: Auditory + Picture + Orthography
Showalter MA Thesis, 2012
Criterion Test
• Two‐way forced‐choice auditory word‐picture matching test
• 48 items: 24 matched, 24 mismatched
• Matched: e.g., see picture of [kita], hear [kita]
• Mismatched: e.g., see picture of [kita], hear [qaʃu]
• Tested only the ability to discriminate among very different words
(e.g., [kita]‐[qaʃu]), not minimal pairs
• Subjects repeated word learning until 90% accuracy reached
Showalter MA Thesis, 2012
Final Test
• Identical to Criterion Test, except:
• Mismatched items were minimal pairs
• e.g., see picture of [kubu], hear [qubu]
[kubu]
[qubu]
matched
mismatched
Showalter MA Thesis, 2012
Experiment 1 Findings
Main effect of item type
sig. (p<.005)
Main effect of group not
sig. (p=.661)
Interaction of item type
and group not sig. (p=.867)
Showalter MA Thesis, 2012
0000
Experiment 1 Conclusions
• Native English‐speaking learners did not benefit from the availability
of the Arabic spelled forms when discriminating between /k‐q/
minimal pairs
• Three possible explanations:
• The Arabic script is too visually comple x
• The novel auditory contrast is too difficult for learners to perceive
• Some combination of the above
Showalter MA Thesis, 2012
Experiment 2
• Is the Arabic script too difficult? (see, e.g., Abdelhadi, Ibrahim &
Eviatar, 2011)
• Attempt to moderate this difficulty with explicit instruction about the
Arabic writing system
Showalter MA Thesis, 2012
Experiment 2
• Procedures identical to Experiment 1, except:
• Prior to the word learning phase, learners presented with information about
the Arabic script
• Only the Orthography condition (n=8)
• Subjects were told that Arabic is written from right‐to‐left and not
left‐to‐right like English and were shown examples with arrows
pointing to the distinguishing letters
Showalter MA Thesis, 2012
Experim ent 2 Findings
• Orthography condition only
• Matched items: .90 (E1: .87)
• Mismatched items: .42 (E1: .47)
Showalter MA Thesis, 2012
Experiment 2 Conclusions
• Explicit instruction (of this type) does not appear to have been helpful
• Did instruction add too much new information for English speakers to
interpret?
• New orthography, new contrasts, and instructions to keep in memory
Showalter MA Thesis, 2012
Experiment 3
• Alleviate script difficulty by using Romanized Arabic script
• Procedures the same as Experiment 1, except:
• Orthographic forms presented in the Roman alphabet (e.g., <kashu>, <qashu>)
• But note novel grapheme‐phoneme correspondence : <q>‐/q/
• Only the Orthography condition (n=8)
Showalter MA Thesis, 2012
Experim ent 3 Findings
• Orthography condition only
• Matched items: .84
(E1: .87, E2: .90)
• Mismatched items: .26
(E1: .47, E2: .42)
Showalter MA Thesis, 2012
Experiment 3 Conclusions
• Difficulty associating with the novel Arabic orthography did not fully
explain native English speakers’ performance in Experiment 1
• Less accurate performance in Experiment 3
• The auditory/k‐q/ contrast may be too difficult
Showalter MA Thesis, 2012
Experiment 4
• Alleviate auditory contrast difficulty by reducing the number of talk ers
to one
• In pilot studies, word learning was enhanced by reducing number of talkers
• Procedures the same as Experiment 1, except:
• Auditory stimuli produced by only one talker
• Orthography condition (n=15)
• Control condition (n=15)
Showalter & Hayes‐Harb, in prep
• Orthography condition
0.8
• Matched items: .87
0.7
• Mismatched items: .62
0.6
(performance sig. above
0.5
chance; p=.037)
0.4
• Control condition
0.3
• Matched items: .89
0.2
• Mismatched items: .52
0.1
0
1
Orthography
0.87
Experiment 4 Findings
0.89
0.9
Control
0.62
0.52
Matched Mismatched
Showalter & Hayes‐Harb, in prep
Showalter & Hayes‐Harb, in prep
D‐prime values per experiment
Conclusions
• Arabic script may be too difficult for learners to utilize
• How is performance accuracy on Experiment 3 (Roman representations)
explained?
• Is the /k/‐/q/ contrast too difficult?
• Even with the aid of orthographic representations?
• Easier contrast?
General Conclusions:
Role of orthographic input in L2 word learning
• Orthographic information can be a powerful source of information
about the phonological structure of L2 words
• Orthographic information may “override” auditory information when…
• Auditory and orthographic information “conflict” (Hayes‐Harb et al. 2010); and
• Auditory information is unusable because learners cannot perceive the
auditory contrast (Escudero et al. 2008)
• Orthographic input may cause learners to…
• Misremember the phonological forms of newly‐learned L2 words
(Hayes‐Harb et al. 2010); and
• Create non‐target‐like phonological representations for L2 syllables
(Bassett i 2006)
General Conclusions:
Influence of orthographic familiarity in moderating role
of orthographic input
• Orthographic information can help learners to associate novel
phonological contrasts with L2 words when…
• The letters are familiar (Dutch/English; Escudero et al., 2008); and
• When the orthographic forms involve unfamiliar symbols (English/Mandarin;
Showalter & Hayes‐Harb, 2013)
• However, orthographic information may not always be helpful to the
acquisition of target‐like L2 forms
• It may have no significant beneficial eff ect when…
• The orthographic and/or auditory contrasts are too difficult for learners to
perceive (Showalter 2012)
• When a novel script is utilized‐‐how do learners interpret a novel script?
General Conclusions:
Talker variability in novel word learning
• Both Showalter and Hayes‐Harb (2013) and Showalter (2012) found
that reduction in number of speakers aided performance
• Pilot studies indicate that L2 word form learning may be more difficult
when input comes from multiple talk ers
• Note: contrary to benefits of talk er variability attested in other domains of L2
phonological acquisition (e.g., work by Pisoni et al., Barcroft and Sommers
2005)
Acknowledgments
• Rachel Hayes‐Harb
• Jane Hacking
• Aaron Kaplan
• Brian Cragun
• Speech Acquisition Lab at the University of Utah
Thank you.
