Introduction Background Data Origins Synchronic implications Conclusions

Gradient Trends against PhoneticNaturalness: The Case of Tarma Quechua

Gasper Begus1 Aleksei Nazarov2

1Harvard Universitybegus@fas.harvard.edu

2University of Huddersfielda.nazarov@hud.ac.uk

NELS 48University of Iceland

October 27-29, 2017

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 1 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Gradient phonotactics

Two aspects of OT widely discussed:

How to represent gradient phonotactic restrictions(Frisch et al. 2004, Antilla 2008, Coetzee and Pater 2008, Wilson and

Obdeyn 2009)

How to represent unnatural processes(Hayes 1999, Buckley 2000, Hyman 2001, Blevins 2004, 2008, Yu 2004,

Wilson 2006, Hale and Reiss 2008, Coetzee and Pretorius 2010, Becker et

al. 2011, White 2013, Hayes and White 2013)

No systematic treatment of the intersection: unnaturalgradient phonotactics

Can gradient phonotactic restrictions operate in thephonetically unnatural direction?

Tarma Quechua stop voicing

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 2 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Naturalness

A new division of naturalness

Phonetic tendencies areenforced by contradicted by

natural processes 3 7

unmotivated processes 7 7

unnatural processes 7 3

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 3 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Literature so far

Unnatural categorical process: post-nasal devoicing, confirmedin Tswana with wug-tests (Coetzee and Pretorius 2010)

Most other processes discussed are in fact unmotivated

Some processes labeled as “unnatural” in Hayes and White(2013)

“No [T,D] before stressless rounded vowels”

*

+COR+cont−strid

[−stress+round

]“No [Z] before stressed vowel + obstruent”

*

+cont+voice−ant

[+stress][−son

]

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 4 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Outline

1 Introduction

2 Background

3 Data

4 Origins

5 Synchronic implications

6 Conclusions

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 5 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Tarma

Tarma Quechua a dialect of Quechua spoken in Tarma, Junın,Peru (Adelaar 1977, Puente Baldoceda 1977)

Distribution of [±voice] in [DOR] and [LAB] stops

Adelaar (1977): [+voice]: intervocalically, post-consonantally,but not post-nasally

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 6 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

From Adelaar (1977):

b, g / C ; C 6= N

b, g / V V

p, k / elsewhere

# [pirwa]R, T [rikra]

N [wampu]V V [kuba]R, T [takba]

Adelaar (1977) offers no further descriptions on thedistribution

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 7 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

From Adelaar (1977):

b, g / C ; C 6= N

b, g / V V

p, k / elsewhere

# [pirwa]R, T [rikra]

N [wampu]V V [kuba]R, T [takba]

Adelaar (1977) offers no further descriptions on thedistribution

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 7 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

From Adelaar (1977):

b, g / C ; C 6= N

b, g / V V

p, k / elsewhere

# [pirwa]R, T [rikra]

N [wampu]V V [kuba]R, T [takba]

Adelaar (1977) offers no further descriptions on thedistribution

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 7 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

Lexical, phonetic, and morphophonological analysis

Unnatural gradient phonotactic restrictions

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 8 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

Distribution of voicing

Native vocabulary from Adelaar (1977)

Counts:

All tokens with [DOR] or [LAB] in Tarma Quechua vocabulary(Adelaar 1977)1199 tokens: 910 in native vocabulary, 289 in loans fromSpanishEach data point was annotated for presence or absence ofvoicing, place of articulation of the stop (labial or velar), andposition in the word

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 9 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

Counts:

# N V V R T

voiced 7 7 99 72 68voiceless 276 67 134 13 11% voiced 2.5 9.5 42.5 84.7 86.1

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 10 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

Logistic regression model:

Est. SE z value Pr(>|z|)(Intercept) -0.045 0.172 -0.260 0.7952V V vs. R 2.044 0.332 6.164 0.0000V V vs. T 2.155 0.353 6.101 0.0000V V vs. N -1.884 0.421 -4.478 0.0000V V vs. # -3.437 0.407 -8.437 0.0000

velar vs. labial -0.502 0.214 -2.344 0.0191

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 11 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

0

25

50

75

# N V V R T

Position

%vo

iced

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 12 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

0

25

50

75

# N V V R T

Position

%vo

iced

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 13 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

0

25

50

75

# N V V R T

Position

%vo

iced

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 14 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

Universal tendencies for [+voice] Observed significant trends in TQ

T < V V V V < TT < N N < V V < T

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 15 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

Another locus of unnaturalness: TT sequences

1st member 2nd memberLabial Velar

t lutbi mutgiÙ / aÙga>úù a

>úùba ma>úùga

k takba /s Ùasbu

>úùasgiS kaSbi iSgix saxbi manexax-gunasl Ùilbi Ùilgir karba arguj ajba ajgaw kawbu awgis

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 16 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

Second-element stops (labial and velar) are significantly morefrequently voiced (as opposed to voiceless) in clusters with avoiceless first element in Tarma Quechua native vocabulary(β = 1.8, z = 5.6, p < 0.0001)

TT TD DT DDCount 11 68 0 0Percent 13.9% 86.1% 0% 0%

All effects thus far remain even if we add loanwords to themodels

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 17 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

Phonetic analysis

Recordings by Willem Adelaar, analyzed in Praat (Boersmaand Weenink 2015)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 18 / 53

Introduction Background Data Origins Synchronic implications Conclusions

[atbi]

749.5

atbi

Time (s)

18.81 19.040

5000

Fre

quen

cy (

Hz)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 19 / 53

Introduction Background Data Origins Synchronic implications Conclusions

[akba]

akba

Time (s)

88.52 88.780

5000

Fre

quen

cy (

Hz)

129.80

5000

Fre

quen

cy (

Hz)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 20 / 53

Introduction Background Data Origins Synchronic implications Conclusions

[ukba]

88.78

ukba

Time (s)

129.8 130.10

5000

Fre

quen

cy (

Hz)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 21 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

After fricatives

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 22 / 53

Introduction Background Data Origins Synchronic implications Conclusions

[asba]

asba

Time (s)

153.2 153.50

5000

Fre

quen

cy (

Hz)

603.70

5000

Fre

quen

cy (

Hz)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 23 / 53

Introduction Background Data Origins Synchronic implications Conclusions

[asga]

153.5

asga

Time (s)

603.7 603.90

5000

Fre

quen

cy (

Hz)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 24 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Data

After nasals

Unaspirated

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 25 / 53

Introduction Background Data Origins Synchronic implications Conclusions

[ampa]

ampa

Time (s)

42.17 42.390

5000

Fre

quen

cy (

Hz)

112.50

5000

Fre

quen

cy (

Hz)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 26 / 53

Introduction Background Data Origins Synchronic implications Conclusions

[aNki]

42.39

anki

Time (s)

112.5 112.90

5000

Fre

quen

cy (

Hz)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 27 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Productivity

Alternating suffixes

-ba/-pa ‘genitive’-bax/-pax ‘purposive’-bita/-pita ‘procedentive’-bis/-pis ‘even, too’

Intervocalicwawxi-gi-ba wayi-n‘the house of your brother’Post-nasalwayi-n-pa pasa-un‘we’re going to walk by way of his house’Post-obstruenttamya-ya-n nuqa-ntik-baq‘it is raining now for us’ (Creider 1968:12-13)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 28 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Productivity

Alternating suffixes

-ba/-pa ‘genitive’-bax/-pax ‘purposive’-bita/-pita ‘procedentive’-bis/-pis ‘even, too’

Intervocalicwawxi-gi-ba wayi-n‘the house of your brother’Post-nasalwayi-n-pa pasa-un‘we’re going to walk by way of his house’Post-obstruenttamya-ya-n nuqa-ntik-baq‘it is raining now for us’ (Creider 1968:12-13)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 28 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Productivity

Alternating suffixes

-ba/-pa ‘genitive’-bax/-pax ‘purposive’-bita/-pita ‘procedentive’-bis/-pis ‘even, too’

Intervocalicwawxi-gi-ba wayi-n‘the house of your brother’Post-nasalwayi-n-pa pasa-un‘we’re going to walk by way of his house’Post-obstruenttamya-ya-n nuqa-ntik-baq‘it is raining now for us’ (Creider 1968:12-13)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 28 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Productivity

Alternating suffixes

-ba/-pa ‘genitive’-bax/-pax ‘purposive’-bita/-pita ‘procedentive’-bis/-pis ‘even, too’

Intervocalicwawxi-gi-ba wayi-n‘the house of your brother’Post-nasalwayi-n-pa pasa-un‘we’re going to walk by way of his house’Post-obstruenttamya-ya-n nuqa-ntik-baq‘it is raining now for us’ (Creider 1968:12-13)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 28 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Productivity

Alternating suffixes

-ba/-pa ‘genitive’-bax/-pax ‘purposive’-bita/-pita ‘procedentive’-bis/-pis ‘even, too’

Intervocalicwawxi-gi-ba wayi-n‘the house of your brother’Post-nasalwayi-n-pa pasa-un‘we’re going to walk by way of his house’Post-obstruenttamya-ya-n nuqa-ntik-baq‘it is raining now for us’ (Creider 1968:12-13)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 28 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Productivity

Loanwords:

Sp. cuculi > kuguli: ‘white-winged dove’Sp. cotpe > kutbi ‘an animal from the mountains’Sp. sauco > sawgu ‘magic tree’Sp. vaca > wa:ga ‘cow’

In two loanwords, a Spanish voiced intervocalic stop devoicesto a Tarma Quechua voiceless stop (data from Adelaar 1977).

Sp. taruga > taruka ‘deer’Sp. dios se lo pague > jusulpa:ki ‘thank you’

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 29 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Outline

1 Introduction

2 Background

3 Data

4 Origins

5 Synchronic implications

6 Conclusions

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 30 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Origins of Tarma Quechua stop voicing

How did this phonotactic restriction arise?

Context Voicing Labial VelarPre-TQ TQ Pre-TQ TQ

# 7 *pirwa pirwa *kawa kawaN 7 *wampu wampu *ÙiNka ÙiNka

V V 3 *kupa kuba *Ùaki ÙagiR,T 3 *takpa takba *kuÙka kuÙga

The most intriguing aspect about this hypothetical soundchange is that this unnatural voicing operates gradientlyrather than categorically with different rates of applicationacross different environments.

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 31 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Origins of Tarma Quechua stop voicing

A diachronic device for explaining unnatural processes:Blurring Process (Begus 2016)

A > B / X natural

B > A / X unnatural

a. A set of segments enters complementary distributionb. A sound change occurs that operates on the

changed/unchanged subset of those segmentsc. Another sound change occurs that blurs the original

complementary distribution

Blurring CycleB > C / −XB > AC > B

Blurring ChainB > C / XC > DD > A

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 32 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Origins of Tarma Quechua stop voicing

Blurring Chain in Tarma Quechua

T > S / [-nas,-#] p > F / [-nas,-#]S > Z / V F > B / VZ > D B > b

Blurring Chain in Tarma Quechua

# V V N T1. pirwa kupa wampu takpa2. pirwa kuFa wampu takFa3. pirwa kuBa wampu takBa4. pirwa kuba wampu takba

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 33 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Distribution

0

25

50

75

# N V V R T

Position

%vo

iced

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 34 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Origins of Tarma Quechua stop voicing

Support from dialectal data:

Fricativization of voiceless stops in Cusco Quechua

*aptay > [haxwtay]*upyay > [uxyay]

Aspiration and fricativization in Imbabura Quechua

Proto-Quechua *paki > *phaki > Imbabura Quechua [faki]Proto Quechua *qipa > *khipa > Imbabura Quechua [xipa]

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 35 / 53

Introduction Background Data Origins Synchronic implications Conclusions

[ubi]

ubi

Time (s)

14.71 14.810

5000

Fre

quen

cy (

Hz)

886.70

5000

Fre

quen

cy (

Hz)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 36 / 53

Introduction Background Data Origins Synchronic implications Conclusions

[atbi]

14.81

atbi

Time (s)

886.7 8870

5000

Fre

quen

cy (

Hz)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 37 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Outline

1 Introduction

2 Background

3 Data

4 Origins

5 Synchronic implications

6 Conclusions

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 38 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Synchronic implications

Deriving typology one of the main advantages of OT

Harmonic Grammar (HG) with numerically weightedconstraints well-suited for gradient processes (Pater 2009)

MaxEnt: Probability distribution over candidates (Goldwaterand Johnson 2003)

Problem that HG approach faces: the derivation of unnaturalprocesses

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 39 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Synchronic implications

OT with restricted Con: factorial typology, unnaturalprocesses unattested (a desired prediction for final voicing)

HG: An additional aspect of the predictive power of HG underthe restricted Con hypothesis that has gone largely unnoticedin the literature:

Natural elements in a given environment will always be morefrequent than unnatural ones

If we allow only natural constraints into Con, we can onlyderive systems with gradient phonotactic restrictions in whichthe natural element in a given context is more frequent thanthe unnatural element

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 40 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Synchronic implications

OT with restricted Con: factorial typology, unnaturalprocesses unattested (a desired prediction for final voicing)

HG: An additional aspect of the predictive power of HG underthe restricted Con hypothesis that has gone largely unnoticedin the literature:

Natural elements in a given environment will always be morefrequent than unnatural ones

If we allow only natural constraints into Con, we can onlyderive systems with gradient phonotactic restrictions in whichthe natural element in a given context is more frequent thanthe unnatural element

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 40 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Synchronic implications

OT with restricted Con: factorial typology, unnaturalprocesses unattested (a desired prediction for final voicing)

HG: An additional aspect of the predictive power of HG underthe restricted Con hypothesis that has gone largely unnoticedin the literature:

Natural elements in a given environment will always be morefrequent than unnatural ones

If we allow only natural constraints into Con, we can onlyderive systems with gradient phonotactic restrictions in whichthe natural element in a given context is more frequent thanthe unnatural element

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 40 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Synchronic implications

E.g. Final Voicing

Restr. Con: *D# 3 *T# 7

Let us assume that all inputs have a uniform prior probability

HG: P(/T#/) = P(/D#/) = 0.5

If the faithfulness constraint (F) Ident-IO(voi) has a positiveinfinite weight and the markedness constraint (M) *D# has afinite weight, the phonotactic probabilities of [T#] and [D#](P([T#] and P([D#]) are both 0.5.

If the markedness constraint is weighted finitely lower than, oreven higher than the faithfulness constraint, the phonotacticprobability of [T#] will be greater than that of [D#]

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 41 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Synchronic implications

With restricted Con, no weighting exists that would yield asystem in which the unnatural feature value has a greaterposterior probability than the natural one in a given context

w(F)− w(M) =∞: P(nat) = P(unnat) = 0.5w(F)− w(M) <∞: P(nat) > P(unnat)

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 42 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Synchronic implications

Natural Gradience Bias (NGB)

HG with restricted Con predicts that the probability of the naturalfeature value in a given environment is always equal or grater thanthe probability of the unnatural value in a given environment.

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 43 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Synchronic implications

NGB correctly predicts the major typological trend with regardto gradient phonotactic restrictions: all cases reportedpreviously indeed operate in the natural direction

As trends in the lexicon, e.g., Berkley 2000, Pater and Coetzee2008, Anttila 2008As tacit phonotactic knowledge obtained from experiments,e.g. Albright 2009

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 44 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Synchronic implications

However, the Tarma Quechua systems of stop voicingpresented in this paper suggest that HG with restricted Conundergenerates

The natural constraints *NT and *T[-voice] will not be ableto give [NT] a higher probability than [ND], or [TD] a higherprobability than [TT]

This, in turn suggests, that Con must contain someunnatural Markedness constraints.

Other such cases: Berawan (Begus and Nazarov 2017)

If we admit all constraints into Con, how to encode rarity ofsome processes?

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 45 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Conclusions and future directions

A case of unnatural gradient phonotactic restriction

Lexical counts, phonetic analysis, signs of productivity

Unnatural gradient phonotactic restrictions find natural origin:Blurring Chain

Synchronic implications: NGB

A challenge to restricted Con

Further experimental work

Other such cases

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 46 / 53

Introduction Background Data Origins Synchronic implications Conclusions

References

Adelaar, Willem F. H. 1977. Tarma Quechua: Grammar, texts, dictionary. Ph.D. dissertation, University of Amsterdam.

Albright, Adam and Youngah Do. 2017.A substantive bias for perceptuallyminimal alternations in Artificial Grammar learning. Presentation at14th Old World Conference on Phonology in Dusseldorf, Germany onFebruary 20-22, 2017.

Begus, Gasper. 2016. Post-Nasal Devoicing and a Probabilistic Model ofPhonological Typology. Ms., Harvard University.

Begus, Gasper and Aleksei Nazarov. 2017. Lexicon against naturalness:Unnatural gradient phonotactic restrictions and their origins. Ms.,Harvard University.

Blevins, Juliette. 2004. Evolutionary Phonology. Cambridge: CUP.Bloyd, Tobias. 2017. Synchronic intervocalic fortition in Sula: a

counter-universal. Presentation at LSA 2017 Annual Meeting in Austin,TX on January 5-8, 2017.

Blust, Robert. 2005. Must sound change be linguistically motivated?Diachronica 22 (2): 219–269.

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 47 / 53

Introduction Background Data Origins Synchronic implications Conclusions

References

Blust, Robert. 2013. The Austronesian Languages. Canberra: Asia-PacificLinguistics.

Blust, Robert. 2016. Austronesian against the world: where the P-map ends.Presentation at the 42nd Berkeley Linguistic Society in Berkeley, CA onFebruary 5-6, 2016.

Canty, Angelo and Brian Ripley. 2016. boot: Bootstrap R (S-Plus) Functions.R package version 1.3-18.

Cathcart, Chundra. 2015. NELS 45: Proceedings of the Forty-Fifth AnnualMeeting of the North East Linguistic Society, edited by Thuy Bui andDeniz zyldz. Volume 1. 145-150. Amherst: GLSA.

Coetzee, Andries W. and Rigardt Pretorius. 2010. Phonetically groundedphonology and sound change: The case of Tswana labial plosives.Journal of Phonetics 38: 404-421.

Davison, A. C. and D. V. Hinkley. 1997. Bootstrap Methods and TheirApplications. Cambridge: CUP.

Dickens, Patrick J. 1984. The History of So-Called Strengthening in Tswana.Journal of African Languages and Linguistics 6:97–125.

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 48 / 53

Introduction Background Data Origins Synchronic implications Conclusions

References

Do, Youngah, Elizabeth Zsiga, and Jonathan Havenhill. 2016. Naturalness andfrequency in implicit phonological learning. Talk presented at the 90thAnnual Meeting of the Linguistic Society of America, Washington, DC,January 7-10.

Efron, Bradley. 1979. Bootstrap Methods: Another Look at the Jackknife. TheAnnals of Statistics 7(1): 1-26.

Hayes, Bruce. 1999. Phonetically-driven phonology: The role of OptimalityTheory and inductive grounding. In Michael Darnell, Edith Moravscik,Michael Noonan, Frederick Newmeyer, and Kathleen Wheatly (eds.),Functionalism and Formalism in Linguistics, Volume I: General Papers,pp. 243-285. Amsterdam: John Benjamins, Amsterdam.

Hyman, Larry M. 2001. The Limits of Phonetic Determinism in Phonology:*NC Revisited. In The Role of Speech Perception in Phonology, editedby Elizabeth Hume and Keith Johnson, 141-186. San Diego, CA:Academic Press.

Janson, Tore (1991/1992). Southern Bantu and Makua. Sprache undGeschichte in Afrika 12/13, 1–44.

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 49 / 53

Introduction Background Data Origins Synchronic implications Conclusions

References

Janssens, Baudouin. 1993. Doubles reflexes consonantiques: quatre etudes surle bantou de zone A (bubi, nen, bafia, ewondo) (Unpublished doctoraldissertation). Universite libre de Bruxelles, Faculte de Philosophie etLettres, Bruxelles.

Kiparsky, Paul. 2006. Amphichronic program vs. Evolutionary Phonology.Theoretical Linguistics 32(2): 217-236.

Kiparsky, Paul. 2008. Universals constrain change, change results in typologicalgeneralizations. In Linguistic universals and language change, edited byJeff Good. 23-53. Oxford: Oxford University Press.

Kummel, Martin. 2007. Konsonantenwandel. Wiesbaden: Reichert.Locke, John. 1983. Phonological Acquisition and Change. New York:

Academic Press.Merrill, John. 2015. Nasalization as a repair for voiced obstruent codas in

Noon. In LSA Annual Meeting Extended Abstracts 6:14-1.Merrill, John. 2016. A Historical Account of the Fula and Sereer Consonant

Mutation and Noun Class Systems. University of California, Berkeley,Ms.

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 50 / 53

Introduction Background Data Origins Synchronic implications Conclusions

References

Moreton, Elliott. 2008. Analytic bias and phonological typology. Phonology25(1): 83127.

Moreton, E., and J. Pater. 2012. Structure and substance inartificial-phonology learning. Part I & Part II. Language and LinguisticCompass 6(11): 686-701 and 702-718.

Picard, Marc. 1994. Principles and methods in historical phonology: FromProto-Algonkian to Arapaho. Montreal: McGill-Queen’s Press-MQUP.

R Core Team 2016. R: A language and environment for statistical computing.R Foundation for Statistical Computing, Vienna, Austria. URLhttps://www.R-project.org/.

Rohlfs, Gerhard. 1949. Lautlehre. Vol. 1 of Historische Grammatik derItalienischen Sprache und ihrer Mundarten. Bern: Francke.

Seidl, Amanda, Eugene Buckley, and Alejandrina Cristi. 2007. Complexitytrumps naturalness. Talk presented at the 81st Annual Meeting of theLinguistic Society of America, Anaheim, CA, January 4-7.

Sole, Maria-Josep, Larry M. Hyman, and Kemmonye C. Monaka. 2010. Moreon Post-Nasal Devoicing: The Case of Shekgalagari. Journal ofPhonetics 38 (4): 299–319.

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 51 / 53

Introduction Background Data Origins Synchronic implications Conclusions

References

Steriade, Donca. 2001. The Phonology of Perceptibility Effects: The P-mapand its Consequences for Constraint Organization. UCLA, Ms.

Wang, William S-Y. 1968. Vowel Features, Paired Variables, and the EnglishVowel Shift. Language 44:695-708.

Wilson, Colin. 2006. Learning Phonology with Substantive Bias: AnExperimental and Computational Study of Velar Palatalization.Cognitive Science 30: 945-982.

White, James. 2017. Accounting for the learnability of saltation inphonological theory: A maximum entropy model with a P-map bias.Language 93(1): 1-36.

Xromov, Albert. 1972. Jagnobskij jazyk. Moskva: Nauka.Yu, Alan C. L. 2004. Explaining final obstruent voicing in Lezgian: Phonetics

and history. Language 80(1): 73-97.Yu, Alan, C. L. 2011. On measuring phonetic precursor robustness: a response

to Moreton. Phonology 28(3): 491-518.

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 52 / 53

Introduction Background Data Origins Synchronic implications Conclusions

Thank you!

* We would like to thank Willem Adelaar, Adam Albright, Gaja Jarosz, Jay

Jasanoff, Sasha Lubotsky, Joe Pater, and Kevin Ryan for their useful comments

on various versions of this work. All mistakes are our own. This research is

partially funded by Mind Brain Behavior interfaculty initiative at Harvard

University.

Gasper Begus, Aleksei Nazarov Harvard University and University of Huddersfield

Gradient Trends against Phonetic Naturalness: The Case of Tarma Quechua 53 / 53