Slide 1
Slide 1 text
Language, Culture, and
Australian Exceptionalism
Claire Bowern • WSC5 :: Davis, Jun 21-23, 2019
[email protected]
Slide 2
Slide 2 text
The Problem
How to account for patterns
History (inertia)
Culture, Society
Linguistic Systems
Physiology, cognition
Acquisition process
Chance
Slide 3
Slide 3 text
“[Australian languages] share a number of striking similarities in their phonology
and in their phonetics, which set them apart from most other languages of the
world.” [Tabain and Harrington 2013]
Surprising Phonological Uniformity in Australia
Little variation in inventory
Similar cognates across the country
Similar (superficial) changes in different subgroups
Slide 4
Slide 4 text
Today
● Discussing the intuition
○ ie, is it true?
● Formulating an explanation
○ ie, why?
Slide 5
Slide 5 text
Do we trust the
intuition?
Slide 6
Slide 6 text
Do we trust the intuition?
Generalizations in the literature about Australian languages tend
to be false
● Bowern (2018); Gasser and Bowern (2014)
● But the inventory claims tend to be true
● sound change not absent (cf. Paman,
Arandic, descriptions of other languages)
● but generalizations about relative uniformity
of inventory are true in general
● caveats are around definitions of ‘rare’ or
‘uncommon’ features
impossible to
test
17%
no
33%
yes
35%
yes, but
15%
impossible to test
no
yes
yes, but
Tested claims
about Aus’n lgs.
Slide 7
Slide 7 text
A typical inventory
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
voicing fricatives laminals
Slide 8
Slide 8 text
Variation in inventory
Source: Gasser & Bowern 2014
Slide 9
Slide 9 text
Cognates across the country
● mara: Nyungar, Kungkari, Malyangapa,
Adnyamathanha, Djabugay, Paakintyi,
Warungu, Yalarnnga, Marrgany, Diyari,
Yulparija, Kariyarra, Jiwarli, Watjarri, ...
● mala: Dyirbal, Gudjal, Mayi-Kutuna
● maʔa: Umpila
● maʔ: Wik-Mungkan, Kuku Nganhcara
● ara: Ikarrangal,
● aʔa: Linngithigh,
● mada: Wonarua
● mar: Koko Bera, Kok Nar,
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Languages with *mara 'hand'
Source: Bowern (2016)
Slide 10
Slide 10 text
Maybe
Pama-Nyungan is just a
young family?
Slide 11
Slide 11 text
How old is the family?
Mean: 5,671 years, 4455-6966 95% HPD
Bouckaert, Bowern,
and Atkinson (2018)
Slide 12
Slide 12 text
Why?
Slide 13
Slide 13 text
Why?
Possible reasons:
Not much change [stability]
Lots of change [convergence]
History (inertia)
Culture, Society
Linguistic Systems
Physiology, cognition
Acquisition process
Chance
Slide 14
Slide 14 text
Not much
variation/
change?
Slide 15
Slide 15 text
Variation
Maybe...
● language isn’t as variable (variation leads to change, but with less
variation, there’s less change).
● variation doesn’t play the same role in communication.
● the phonological systems are stable, so the areas we see the most
change in in other languages doesn’t apply here.
● there’s variation, but it’s not under selection, so change is slow.
Slide 16
Slide 16 text
Variation
But ...
● No evidence that there’s less
variation in general; positive
evidence for variation in some
domains.
● pitch peak realization [e.g. Round 2009]
● Prosody-stress relations [Fletcher &
Butcher 2010]
● stop voicing [e.g. Kakadelis 2018]
● grammar descriptive studies
Maybe...
● The signal isn’t as variable (variation
leads to change, but with less
variation, there’s less change).
● Maybe variation doesn’t play the same
role in communication.
● The phonological systems are stable,
so the areas we see the most change in
in other languages doesn’t apply here.
● There’s variation, but it’s not under
selection, so change is slow.
Slide 17
Slide 17 text
140
150
160
170
1 2 3 4 5 6 7 8 9 10
Time (normalized)
F0 (Hz)
Type
init-contested
init-stressed
long
unstressed
Type
Mean Intensity (dB)
50
60
70
80
Initial Long Stressed Unstressed
a
Initial Long Stressed Unstressed
i
Initial Long Stressed Unstressed
u
Evidence for variation
Slide 18
Slide 18 text
Variation
But ...
● Positive evidence for variation that’s
socially conditioned
● Bowern (2007) – Yan-nhaŋu
● Sommer (2003) – Cape York
● Meakins (much work) - Gurindji
● Douglas/Miller exchange – Wati
● Mansfield (2014) - Murrinh Patha
● Yolŋu folk linguistics
● Much informal evidence from speakers
● But little formal study
Maybe...
● The signal isn’t as variable (variation
leads to change, but with less
variation, there’s less change).
● Maybe variation doesn’t play the
same role in communication.
● The phonological systems are stable,
so the areas we see the most
change in in other languages doesn’t
apply here.
● There’s variation, but it’s not under
selection, so change is slow.
Slide 19
Slide 19 text
Variation
No one’s investigated this (yet)...
● Babinski and Bowern (2018) replicated
Wedel et al’s (2013) findings for Bardi.
● More mergers in areas of lower functional load.
● Differential distribution of minimal pairs over
different languages
● No systemic discussion of which sounds are
more likely to change
○ (data only recently available for worldwide study)
○ not clear that this can be done with current
transcriptions
Maybe...
● The phonological systems are
stable, and the areas where we
see sound change in other
languages aren’t applicable.
● Even functional load across system
might impede change
● Inventories might be disproportionately
composed of stable segments
Slide 20
Slide 20 text
Variation
But ...
● There’s evidence for selection.
○ Production biases apply in Australia as elsewhere
(as far as studied).
○ social selection (cf. below)
○ Uniformitarian principle argues against this too.
● Where there is change, it looks like we find
elsewhere (cf. Bardi; Babinski & Bowern 2018;
Koch 2007, Alpher 2004, etc)
○ Lenition more common than fortition
○ Final devoicing
○ Palatalization
○ Voicing / vowel length compensatory changes
Maybe...
● The signal isn’t as variable (variation
leads to change, but with less variation,
there’s less change).
● Maybe variation doesn’t play the same
role in communication.
● The phonological systems are stable, so
the areas we see the most change in in
other languages doesn’t apply here.
● There’s variation, but it’s not under
selection, so change is slow.
Slide 21
Slide 21 text
Too much
change?
Transmission
and contact
Slide 22
Slide 22 text
Transmission
Maybe...
● Transmission is different: highly constrained social networks of non-interaction (e.g. can’t
talk to opposite sex relatives, mothers-in-law) make it hard for innovations to spread.
● Dense/multiplex social networks might impede diffusion of change, make transmission
more faithful
● There are physiological bottlenecks [e.g. Otitis media]
● Multilingual communities lead to intensive contact and convergence
● Maybe there is sound change but it’s obscured by high rates of lexical replacement.
Slide 23
Slide 23 text
Transmission
But...
● We see innovation in other areas of
culture (and language).
● Others have deduced Australia’s
exceptional fluidity and propensity to
language change!
● The social network interactions are
complex; would lead to more
fragmentation and variation?
● The idea of Australian stasis is a
myth.
Maybe...
● Transmission is different: highly
constrained social networks of non-
interaction (e.g. can’t talk to opposite sex
relatives, mothers-in-law) make it hard for
innovations to spread.
● Dense/multiplex social networks might
impede diffusion of change, make
transmission more faithful
Slide 24
Slide 24 text
Transmission
But...
● Australian social networks also
contain ‘weak’ links
○ e.g. Wati
○ exogamy ensures regional ties
● Dense social networks also lead to
faster change when innovations do
arise.
● Australian groups aren’t uniform
here! (can’t use as global
explanation)
Maybe...
● Transmission is different: highly constrained
social networks of non-interaction (e.g. can’t
talk to opposite sex relatives, mothers-in-law)
make it hard for innovations to spread.
● Dense/multiplex social networks might
impede diffusion of change, make
transmission more faithful (cf.
Granovetter 1983 and much recent work)
Slide 25
Slide 25 text
Transmission
But...
● Bowern et al (2011) showed that
Australian languages weren’t
different from the rest of the world
● Contact isn’t biased towards feature
conservation
Maybe...
● Multilingual communities lead to
intensive contact and convergence
● There are physiological bottlenecks [e.g.
Otitis media]
● Maybe there is sound change but it’s
obscured by high rates of lexical
replacement.
Slide 26
Slide 26 text
Transmission
But...
● Time scale?
● Not clear community rates are not
inflated by colonial practices
(missions, inappropriate housing
without access to clean water, etc.)
● Should have coastal gradient.
● Doesn’t explain acquisition of
allophonic variation of sounds that
are hard to perceive.
● Cf. Fergus (2019): no evidence
Maybe...
● Multilingual communities lead to intensive
contact and convergence
● There are physiological bottlenecks [e.g.
Otitis media – idea due to Andy Butcher]
○ Chronic middle ear infections create
transmission bias against acquisition of stop-
fricative contrasts
● Maybe there is sound change but it’s
obscured by high rates of lexical
replacement.
Slide 27
Slide 27 text
Transmission
But...
● Death taboos vary across the
country.
● Most don’t last long, and don’t affect
the whole community (they don’t lead
to lexical replacement mostly)
● It’s not clear that the replacement
rates are higher than elsewhere.
Maybe...
● Multilingual communities lead to intensive
contact and convergence
● There are physiological bottlenecks [e.g.
Otitis media]
● Maybe there is sound change but it’s
obscured by high rates of lexical
replacement (e.g. due to death taboos).
Slide 28
Slide 28 text
Conclusions
Slide 29
Slide 29 text
Conclusions
Genuinely exceptional?
Yes, probably.
Not much change?
None of the arguments are convincing, but not all have been properly investigated.
Therefore the areas that haven’t been investigated are the only options at this point [e.g.
inventory structure]
Too much change?
None of the arguments are particularly convincing, but need to look at rates of lexical
replacement w.r.t. other families.
Slide 30
Slide 30 text
Thank you!
BCS-1423711