The Articulation of Scottish Gaelic Plain and Palatalized
ConsonantsJae-Hyun Sung1, Diana Archangeli1,2, Ian Clayton1,3,Dan
Brenner1, Sam Johnston1, Michael Hammond1,and Andrew Carnie1
University of Arizona1; University of Hong Kong2; Boise State
University3
[email protected]; [email protected];
[email protected];{dbrenner, sjcjohnston, hammond,
carnie}@email.arizona.edu
Research QuestionsWhat do articulatory gestures tell us about
palatal/palatalized conso-nants in Scottish Gaelic (henceforth
SG)?
1. How are plain, palatal, and palatalized consonants
different?2. Do liquids have the same articulation whether in plain
or palatal/palatalized
contexts?3. How are different syllabic positions at play? (e.g.,
initial vs. final)4. Is the typical description accurate, that
Lewis Scottish Gaelic /l/ & /r/
has plain and palatal versions? What about Uist?
Data
• Palatalization appears as morphologically-conditioned
consonant mu-tation in SG (Macaulay (1992), Ladefoged et al.
(1998), Gillies (2002),Stewart (2004))
• Both phonemic and morphological
Data 1: Plain vs. Palatal (1) vs. Palatalized (2)
[p] vs. [pj]1 vs. [pj]2 plain go ’beak’ (nom.sg.)palatal slai
’muck’palatalized gui ’beak’s’ (gen.sg.)
[l] vs. [L]1 vs. [L]2 plain Ga ’lowlander’ (nom.sg.)palatal
ainmei ’famous’palatalized Goi ’lowlander’s’ (gen.sg.)
Data 2: Initial vs. Final Palatal/Palatalized
[pj] initial inn ’verdict’final gui ’beak’s’ (gen.sg.)
[kj] initial inn ’grow’final gli ’wise’
Data Collection & Measurements
• 26 speakers of SG recruited in Sabhal Mòr Ostaig, Isle of
Skye• Data from 4 speakers: 3 Lewis speakers & 1 Uist
speaker
– Speaker #7: born in Lewis of parents both born in Lewis–
Speaker #10: born in Uist of parents both born in Uist– Speaker
#15: born in Lewis of parents both born in Lewis– Speaker #25: born
in Lewis of parents both born in Lewis
• 240 tokens per speaker considered for analysis (240 tokens × 4
speak-ers = 960 tokens in total)
• Ultrasound images of gestural peaks selected & traced
manually
ResultsResult 1: Plain (green) vs. Palatal (red) vs. Palatalized
(blue) Non-liquids
1a: Plain vs. Palatal/Palatalized
#7 #10 #15 #25
/p/ −340−2
80−2
20
tong
ue h
eigh
t (pi
xels
) final−p+palfinal−p−pal
−320
−260
−200
tong
ue h
eigh
t (pi
xels
) final−p+palfinal−p−pal
−340
−280
−220
tong
ue h
eigh
t (pi
xels
) final−p+palfinal−p−pal
−340
−300
−260
tong
ue h
eigh
t (pi
xels
) final−p+palfinal−p−pal
/t/ −320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−t+palfinal−t−pal
−320
−260
−200
tong
ue h
eigh
t (pi
xels
) final−t+palfinal−t−pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−t+palfinal−t−pal
−340
−300
−260
tong
ue h
eigh
t (pi
xels
) final−t+palfinal−t−pal
1b: Palatal vs. Palatalized
#7 #10 #15 #25
/t/ −320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−t+palatalfinal−t+palatalized
−320
−260
−200
tong
ue h
eigh
t (pi
xels
) final−t+palatalfinal−t+palatalized
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−t+palatalfinal−t+palatalized
−340
−300
−260
tong
ue h
eigh
t (pi
xels
) final−t+palatalfinal−t+palatalized
Result 2: Plain (green) vs. Palatal (red) vs. Palatalized (blue)
Liquids
2a: Plain vs. Palatal/Palatalized
#7 #10 #15 #25
/l/ −300
−260
−220
tong
ue h
eigh
t (pi
xels
) final−l+palfinal−l−pal
−300
−260
−220
tong
ue h
eigh
t (pi
xels
) final−l+palfinal−l−pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−l+palfinal−l−pal
−300
−260
tong
ue h
eigh
t (pi
xels
) final−l+palfinal−l−pal
/r/ −300
−260
tong
ue h
eigh
t (pi
xels
) final−r+palfinal−r−pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−r+palfinal−r−pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−r+palfinal−r−pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−r+palfinal−r−pal
2b: Palatal vs. Palatalized
#7 #10 #15 #25
/l/ −300
−260
−220
tong
ue h
eigh
t (pi
xels
) final−l+palatalfinal−l+palatalized
−300
−260
−220
tong
ue h
eigh
t (pi
xels
) final−l+palatalfinal−l+palatalized
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−l+palatalfinal−l+palatalized
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) final−l+palatalfinal−l+palatalized
Result 3: Initial (red) vs. Final Palatal/Palatalized
(green)
#7 #10 #15 #25
/p/ −320
−260
−200
tong
ue h
eigh
t (pi
xels
) initial−p+palfinal−p+pal
−300
−250
−200
tong
ue h
eigh
t (pi
xels
) initial−p+palfinal−p+pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) initial−p+palfinal−p+pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) initial−p+palfinal−p+pal
/t/ −320
−280
−240
tong
ue h
eigh
t (pi
xels
) initial−t+palfinal−t+pal
−320
−260
−200
tong
ue h
eigh
t (pi
xels
) initial−t+palfinal−t+pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) initial−t+palfinal−t+pal
−300
−260
tong
ue h
eigh
t (pi
xels
) initial−t+palfinal−t+pal
/l/ −300
−260
−220
tong
ue h
eigh
t (pi
xels
) initial−l+palfinal−l+pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) initial−l+palfinal−l+pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) initial−l+palfinal−l+pal
−300
−260
tong
ue h
eigh
t (pi
xels
) initial−l+palfinal−l+pal
/r/ −300
−260
−220
tong
ue h
eigh
t (pi
xels
) initial−r+palfinal−r+pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) initial−r+palfinal−r+pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) initial−r+palfinal−r+pal
−320
−280
−240
tong
ue h
eigh
t (pi
xels
) initial−r+palfinal−r+pal
DiscussionRQ Comparison ResultsQ1a plain vs. palatal(ized)
mostly differentQ1b palatal vs. palatalized obstruents no
differenceQ2a plain vs. palatal(ized) mostly no differenceQ2b
palatal vs. palatalized liquids mostly differentQ3 initial vs.
final palatal(ized) sometimes differentQ4 dialectal difference not
observed
• Articulatory patterns of palatalization are different across
consonanttypes (Q1a & Q1b).
• Gestural patterns here also capture palatal vs. palatalized
contrast inliquids (Q2b); partially confirms the theoretical claim
on palatal/non-palatal contrast in liquids (Borgstrøm (1941), Lamb
(2003), Ternes (2006),Maolalaigh (2008))
• Different syllabic positions play a greater role in liquids
than in ob-struents (Q3).
• The results do not reflect any dialectal difference between
Lewis andUist speakers (Q4).
Conclusions
• Overall, our preliminary results show a clear sign of
palatalization inpalatal(ized) consonants (i.e., plain vs.
palatal/palatalized).
• Different underlying representations (i.e., palatal vs.
palatalized) andsyllabic positions (i.e., initial vs. final) often
create gestural differences,but do not always manifest as
differences in tongue contours.
Selected References
Borgstrøm, C. H. 1941. The Dialects of Skye and Ross-shire: A
Linguistic Survey of the GaelicDialects of Scotland. H. Aschehoug
& Co.Gillies, W. 2002. Scottish Gaelic. In Ball, M. and Fife,
J., editors, The Celtic Languages,pages 145–227.
Routledge.Ladefoged, P., Ladefoged, J., Turk, A., Hind, K., and
Skilton, S. J. 1998. Phonetic struc-tures of Scottish Gaelic.
Journal of the International Phonetic Association, 28:1–41.Lamb, W.
2003. Scottish Gaelic. Lincon Europa.Macaulay, D. 1992. The
Scottish Gaelic language. In Macaulay, D., editor, The
CelticLanguages, pages 137–248. Cambridge University
Press.Maolalaigh, R. Ò. 2008. Scottish Gaelic in twelve weeks.
Birlinn.Stewart, T. W. 2004. Mutation as morphology: bases, stems,
and shapes in Scottish Gaelic.PhD thesis, The Ohio State
University.Ternes, E. 2006. The phonemic analysis of Scottish
Gaelic: based on the dialect of Applecross,Ross-shire. Dublin
Institute of Advanced Studies.
AcknowledgementsThe authors gratefully acknowledge support from
the National ScienceFoundation grant to Andrew Carnie (award
#BCS11443818) and JamesS. McDonnell Foundation grant to Diana
Archangeli (award #220020045BBNB).
