Segmental Inventories, GLOW Workshop, 11th April 2007.

1

The Nature of Stop Contrasts Predicts Prenasalisation* Karthik Durvasula University of Delaware [email protected] 0. Introduction
Contributions of the paper: •

Typology of pre/post-nasalised stops (or partially-nasal stops).

•

Analysis within Dimensional Theory (Avery & Idsardi (2001)) – explicit about the phonology-phonetics interface.

0.1 Pre/post-nasalised stops and related phenomena
Segments with both a nasal and an oral portion. a. Pre-nasalised stops (nd, mb, ŋk) - usually in onsets. b. Post-nasalised stops (dn, bm, kŋ) – usually in codas.
(a-b) are claimed not to contrast in languages.; hence, ordering argued to be phonologically irrelevant.
‘Partially-nasal stops’ for the rest of the talk. 0.2 Main Questions and Conclusions I. Are the phonological representations of simple nasal stops and partially-nasal stops different? (c. Chomsky and Halle (1968), Anderson (1976), Sagey (1986), Piggott (1992), Rice (1993), Botma (2004)) •

Claim – No, they have the same phonological representation.

II. What effect does contrast have on the phonetic interpretation/manifestation of phonological representations? (Padgett (2001), Liljencrants & Lindblom (1972), …) • *

Claim – Increased contrast enhances phonological dimensions.

I would like to thank Dr. William Idsardi, Dr. Peter Cole, Dr. Uri Tadmor, Dr. Irene Vogel, Dr. Eric Raimy, and all the students at the Dept. Of Linguistics, University of Delaware for their help and support.

2

Karthik Durvasula

III. Is contrast a global/local phenomenon, i.e., do we ‘calculate’ contrast by looking at segment inventories or do we ‘calculate’ it by looking at local featural oppositions? •

Claim – Contrast calculation is via local featural oppositions.

IV. Does

diminished

phonological

contrast

always

mean

increased

representational/phonological underspecification? •

Claim – No!

1. Typology of partially-nasal stops 1.1 Partially-nasal stops and 2-way stop contrast
Partially-nasal stops typically appear in languages with a 2-way stop contrast (1a) – Barasano, Kaingang, Guarani … (Anderson (1976), Rice (1993), van de Weijer and Hinskens (2004) inter alia). (1)a. 2-way stop contrast Oral Stops Nasal Stops

p

m

b

b. 3-way stop contrast t

n

k d ŋg

Oral stop(vl./asp.) Oral stop(vd.) Simple nasal stops

p t k b d g m n ŋ

1.2 The nasal/oral alternation
Simple nasals appear adjacent to tauto-syllabic nasal vowels while partially-nasal stops appear next to tauto-syllabic oral vowels. (2) Barasano (Data from Piggott and Hulst (1997)) a. mãsã

‘people’

c. mbaŋgo ‘eater’

e. *m v c v (v =vowel, c =cons.)

b. mãnõ

‘none’

d. ndiro

f. * nc v c v (v = nasal vowel)

‘fly’

1.3 Partially-nasal stops have (voiced) oral stop counterparts in related dialects
Languages with partially-nasal stops sometimes have voiced variants appearing in place of partially-nasal stop segments (in some dialects) - (Slave (Rice (1993)), Southern Barasano1 (Botma (2005))). 1

This free alternation in (3) is limited to non-word-initial positions. Word-initially, the nasalisation is obligatory. (i)a. ndiro (*diro) ‘grasshopper’ b. mbango ~ mbago (*bago) ‘eater’

Segmental Inventories, GLOW Workshop, 11th April 2007.

3

(3) Southern Barasano (data from Botma (2005)) a. wamba ~ waba ‘come!’ m b. ta boti ~ taboti ‘grass’
The alternation between partially partially-nasal stops and voiced stops has led some researchers (Rice (1993), Botma (2005) amongst others) to conclude that there is an inherent connection between the features for sonorant, voicing and nasality.
However, it will be shown later that a more precise generalization for these dialects is that they vary between oral and partially-nasal stops. 1.4 Inconsistent voicing in partially-nasal stops
Voicing has been claimed to be an important phonetic correlate of partially-nasal stops. Stevens and Keyser (2006) and van de Weijer and Hinskens (2004) – they attribute this to the phonemic representation of these segments

A

B

Fig 1. Broadband spectrogram of the Fijian word - ndanda ‘soft’ (Maddieson (1989)).
Observe the high-lighted portions A & B - If partially-nasal segments were underlyingly specified for voicing, then it is unclear why they would not show a consistent voicing bar throughout.

4

Karthik Durvasula

1.5 Partially-nasal stops do spread nasality
Some researchers (Rice (1993), Botma (2004)) have argued that partially-nasal stops (in most languages) are phonologically not specified for the feature nasal and are different from simple nasals – thereby implying that they cannot spread nasality2. However, partially-nasal stops have been observed to spread nasality in at least three languages – Guaraní (Walker (1998), and references within), Yuhup (Botma (2005) and references within), Tinrin (Botma (2004) and references within).
However, partially-nasal stops have been observed to spread in a few languages. (4) Partially-nasal stops (regressively) spreading nasality a. Guaraní - Walker (1998) /ro + mbo +γwatá/ ––––––> [rõmbo γwatá] ‘I made you walk’ b. Yuhup - Botma (2005) /tə:dn + ih/

––––––> [tə:dnĩh]

‘beating’

c. Tinrin – Botma (2004) /fa + nde/

––––––> [fãnde]

‘hang something up’

1.6 Summary of facts about partially-nasal stops A. Partially-nasal stops (usually) appear in 2-way stop contrast systems. B. In these languages, partially-nasal /oral stops appear next to oral vowels, while simple nasal stops appear next to nasal vowels. C. Some dialects of these languages have (voiced) oral stops in place of partiallynasal stops (i.e, only next to tauto-syllabic oral vowels) – it will be shown in the next section that these dialects show orality as the important characteristic in these stops, not the voicing that generally accompanies them. D. Voicing is not a consistent phonetic correlate of the oral portion of partially-nasal stops. 2

Botma (2004) specifically claims that partially-nasal stops (in most cases) are ‘phonologically inert nasals’, as opposed to simple nasals which spread nasality which he calls ‘phonologically active nasals’. The testability of his theory nearly vanishes when he claims that partially-nasal stops which appear in conditions similar to other languages having them but spread nasality are ‘phonologically active nasals’.

Segmental Inventories, GLOW Workshop, 11th April 2007.

5

E. Partially-nasal stops can and do spread nasality to adjacent segments. (E) strongly motivates the claim that partially-nasal stops are phonologically specified for nasality. 2. Case studies
Jambi Malay and Mamaindé show some segments which show the properties of partially-nasal segments discussed above. Claim: simple nasal stops appear in 3-way (or more) stop contrast environments, and partially-nasal stops (may) in 2-way stop contrast environments. 2.1 Jambi Malay
Tanjung Raden dialect of Jambi Malay - a Malay dialect spoken in the province of Jambi, Sumatra (Tadmor and Yanti (2005)). (5)a. Non-word-finally (4-way stopcontrast)

b.Word-finally(2-way stop contrast)

laba-laba ‘spider’

ika(n)t

‘fish’

bapa?

‘father’

makadn

‘eat’

kamãr

‘room’

tambat3

‘to tie’ (taMat)

(6) Nasal/Oral Alternation4 a. After oral vowels

3

b. After nasalised vowels

it(t)abm

‘black’

minũm

‘drink’

kapadn

‘when’

dəŋãn

‘and’

siagŋ

‘early morning’

bərənãŋ

‘swim’

These word-medial segments (‘mb’) have been described as ‘post-occluded nasals’ (Tadmor and Yanti (2005)). However, they don’t seem to have any ‘oral’ component (similar nasals in Acehnese (Durie (1985)) show no oral portion. Hence, they are possibly inaccurately transcribed. 4 Tadmor and Yanti (2005) argue that pre-occlusion with these nasal sounds is ‘wordfinal’ only in citation form, and is actually utterance final. See note 24 for more on this possibility.

6

Karthik Durvasula


Variation is between voiceless/nasal/partially-nasal stops, but NOT voiced stops! (7) Dialectal (neighbouring language) Variation Tanjung Raden Sarolangun5 a. makadn makat

‘eat’

b.

tulagŋ

tulak

‘bone’

c.

jarubm

jarup

‘needle’


No inherent connection between sonorant/voicing/nasal. 2.2 Mamaindé
Mamaindé – An Amazonian language of west-central Brazil. It belongs to the Nambiquara langauge family (Eberhard (2004)).
Mamaindé has a 3-way stop contrast in the onset position, but a 2-way stop contrast in the coda position6 (data from Eberhard (2004)). (8)a. Onset Position i. ha:.na?.wa ‘I wiggle’ ii. ja.da:.nã.ni ‘the deer’

b. Coda Position i. naik.tu7 ‘a root’ g ii. si ŋ.du ‘a meat’

iii. hadn.la.tha.wa ‘It wiggles’
Furthermore, these partially-nasal stops become simple nasal stops when adjacent to a tauto-syllabic nasal vowel as shown by (9). (9)a. ja.lãn.du ‘toucan’ b. ja.dadn.du 5

‘deer’

This data is from my fieldwork in Jambi, Indonesia during the January 2006. However, more recent fieldwork by Tim McKinnon (University of Delaware) shows Sarolangun to have variation between partially-nasal stops and voiceless stops in word-final position. However, this does not affect the analysis presented here. 6 This was not explicitly mentioned/acknowledged in Eberhard (2004), but he (Eberhard) has confirmed this fact via email to the author, Karthik Durvasula. 7 From Eberhard’s (2004) paper and his email, one is led to believe that voiceless unaspirated stops and voiced unaspirated stops are allophones in this language as evidenced by: (i) [naik.tu] ‘a root’ (ii) [nai.ga.ni] ‘the root’

Segmental Inventories, GLOW Workshop, 11th April 2007.

7


Mamaindé provides strong support to the observation that a 2-way stop contrast is crucial for nasals to surface as partially-nasal stops. (10)a. In the coda position /han + lathawa/

[hadn.la.thwa]

‘it flops’

[ha.ná?.wa]

‘It does not flop’

Stop Contrast

Nasal manifestation

4-way

Simple nasals

Word-final

2-way

Partially-nasal stops

Onset

3-way

Simple nasals

Coda

2-way

Partially-nasal stops

–––––>

b. In the onset position /han + á?wa/

–––––>

2.3 Conclusions of the case-studies (11) a.Jambi Malay Non-final

b.Mamainde


Phonologically, partially-nasal stops and nasals are the same; there is no inherent connection between the features nasal/sonorant/voicing. 3. Recent Proposals 3.1 Proposals8
Sonorant Voicing (SV) Node Analysis (Piggott (1992), Rice (1993)).
Element Theory framework (Botma (2004)).
Common points (a) Nasal

stops

and

partially-nasal

stops

have

different

phonological

representations. (b) Dialectal variants between nasals/voiced/partially-nasal stops are related to the phonetic manifestation of ‘sonorant’. (c) Partially-nasal stops are not capable of spreading nasality.

8

A gestural overlap account of partially-nasal stops (a la Browman and Goldstein (1989)) has been argued against in Appendix I.

8

Karthik Durvasula

3.2 Problems
Partially-nasal stops CAN spread nasality. (section 1.5)
No answer to -Why do ‘sonorant stops’ surface as voiceless stops (rathr than voiced stops) in Sarolangun Malay? (7a-c)
The proposals cannot explain the simple-nasal stop/partially nasal-stop alternations in Mamaindé (section 2.2), except with an alternation specific rule/constraint.
Apart from this, each of the proposals has further specific problems. For eg., the SVnode analysis. (12)

Sonorant Voicing SV

Lateral (Nasal) (where, ‘nasal’ is the unmarked sonorant, hence in parentheses)
(13a) shows the representations for the different kinds of nasal stops, where (13b) is phonologically unspecified for any terminal node. (13)a. Simple Nasal Stop Root | SV | [nasal]

b. Simple nasal/partially-nasal stop/(voiced)oral stop Root | SV


The representation in (13b) can surface as ‘simple nasals’, ‘partially nasal stops’ or ‘voiced stops’. This seems to be giving quite some latitude to the phonetic component. What would be the phonetic representations of such segments? Such a degree of latitude may be unnecessary. 4. New Analysis 4.1 Dimensional Theory - Avery and Idsardi (2001)
The terminal features are all what Avery and Idsardi call gestures - motor instructions to the articulators. The pre-terminal nodes are called dimensions, which are abstract organizations/nodes that pair phonetically-antagonistic “gestures”.

Segmental Inventories, GLOW Workshop, 11th April 2007.

9

(14)


In Avery and Idsardi’s model, phonological representations are specified only to the degree that is necessary to maintain contrast within the phonological system.
Furthermore, representations are either marked for a dimension or not – essentially, Ø/Marked.
The dimensions are realised phonetically through the use of completion rules that supply the actual phonetic gesture that will be implemented under the dimension – (15a) represents one type of laryngeal system possible in the proposed theory (English is argued to be such a system); in (15b), the dimension GW is completed with a [spread (vocal folds)] gesture. b. Phonetic Implementation9

(15) a. Phonological Representation i. V.less aspirated stops Root | GW

9

ii. Vd. Stops Root

i.V.less aspirated stops Root | GW | [spread]

In a system such as the one represented in (15), the unmarked segment is claimed to vary in its phonetic manifestation (in this case, only for the laryngeal features), and for that reason no phonetic representation is shown for the voiced stop in (15a-ii)

10

Karthik Durvasula


Another important characteristic of their analysis is that they allow for contextual completion of different dimensions. For example, English voiceless stops are aspirated in the onset but glottalised in the coda. The GW dimension is completed with [spread glottis] in the onset and [constricted glottis] in the coda – this depends on the phasing relationship between the dimension GW and the oral constriction. 4.2 Formal analysis for nasals and partially-nasal stops
The relevant dimension/organizational node: (16)

SP (Soft Palate) / \ [oral] [nasal]

(17) Phonetic implementation of the Soft Palate (SP) node a. Simple Nasals X | SP | [nasal]

b. Oral stops X | SP | [oral]

c. Partially-Nasal Stops X | SP


Both simple nasals and partially-nasal stops are phonologically represented in the same way as in (18). (18) Phonological representation of simple nasals / partially-nasal stops10 X (root node) | SP
The above phonological representation (18) trivially derives the fact that partially nasalised segments can spread nasality as they are specified for the SP dimension.
The gesture (nasal/oral/ Ø) with which the dimension ‘SP’ is completed/implemented is dependent on context.

10

These representations and those that follow, are of course, partial representations of the segments being discussed.

Segmental Inventories, GLOW Workshop, 11th April 2007.

11

(19) Phonetic rule of completion for SP. Nasal Completion Rule: In contexts with more than a 2-way stop contrast, the SP node is necessarily completed with the gesture [nasal].
The completion rule in (19) can be seen as a rule of enhancement (Keyser and Stevens (2006), and references therein) of the dimension SP in the face of an increased number of contrasts.
The completion rule in (19) derives the following: a. Languages with 3-way stop systems will have phonological nasals that surface as simple nasals (20a). b. Languages with a 2-way stop system are expected to show dialectal variation in the actual phonetic manifestation of the SP dimension as there is no pressure on the dimension to be completed/enhanced with nasal gesture (20b). Phonological Reps. (20)a. 3-way stop contrast

X | SP

b. 2-way stop contrast

X | SP

Phonetic Reps. X | SP | [nasal] X | SP | [nasal]/[oral]/ [Ø]

5. Conclusion
Partially-nasal stops and simple nasal stops are phonologically identical.
The phonetic differences are due to differing contrast systems, and consequently differing phonetic implementation of the phonological representation.

12

Karthik Durvasula

6. Implications
This paper shows that contrast is central to the phonetic manifestation of (abstract) phonological entities.
The proposal supports the view that ‘contrast’ can and should be defined locally, i.e., not through segment inventories.
It shows that diminished phonological contrast need not indicate a more underspecified segment, as it is not the case that final nasals in Jambi Malay (and coda nasals in Mamaindé) are more underspecified than simple nasals.

References Avery, Peter and William Idsardi (2001) Laryngeal Dimensions, Completion and Enhancement. in T. Alan Hall (ed.), Distinctive Feature Theory, de Gruyter, Berlin, 41-70. Botma, Bert (2004). Phonological Aspects of Nasality: an Element-based Dependency Account. Doctoral dissertation, University of Amsterdam. Botma, Bert (2005). Nasal harmony in Yuhup: a typological anomaly?. In Nancy Kula & Jeroen van de Weijer (eds.) Papers in Government Phonology. Special issue of Leiden Papers in Linguistics 2.4 (2005), 1-21. Chomsky, Noam and Morris Halle (1968). The sound pattern of English. New York: Harper and Row. Browman, Catherine and Louis Goldstein (1989). Articulatory gestures as phonological units. Phonology 6: 201-251. Downing, Laura (2005). On the ambiguous segmental status of nasals in homorganic NC sequences. In M. van Oostendorp and J.M. van de Weijer (eds.) The internal organization of phonological segments. Berlin: Mouton de Gruyter. Durie, Mark (1985). A Grammar of Acehnese on the Basis of a Dialect of North Aceh. Verhandelingen van het Koninklijk Instituut voor Taal- Land- en Volkenkunde Nr. 112. Dordrecht: Foris. Eberhard, David (2004). Mamaindé pre-stopped nasals: an optimality account of vowel dominance and a proposal for the Identical Rhyme Constraint. . Keyser, Samuel Jay and Kenneth Stevens (2006). Enhancement and Overlap in the Speech Chain. Language 82. Liljencrants, J. and Lindblom, B. (1972). Numerical simulation of vowel quality systems: the role of perceptual contrast. Language 48.4: 839-862. Maddieson, Ian (1989). Prenasalized stop and speech timing. Journal of International Phonetic Association. 19(2), 52-66. Nasukawa, Kuniya (2005). A Unified Approach to Nasality and Voicing. Studies in Generative Grammar 65. Mouton de Gruyter

Segmental Inventories, GLOW Workshop, 11th April 2007.

13

Padgett, Jaye (2001). Contrast dispersion and Russian palatalization. In The role of speech perception in phonology, ed. by Hume, E. and Johnson, K.: 187-218. Academic Press, San Diego, CA). Piggott, Glynne (1992). Variability in feature dependency: The case of nasality. Natural Language and Linguistic Theory 10.33-77 Piggott, Glynne (1996). Implications of consonant nasalization for a theory of harmony. Canadian Journal of Linguistics 41. 141–174. Piggott and Hulst (1997) Locality and the nature of harmony. Lingua 103, 85-112 Rice, Keren (1993). A reexamination of the feature [sonorant]: the status of 'sonorant obstruents.' Language 69. 308-344. Rice, Keren and Peter Avery (1989). On the interaction between sonorancy and laterality. Toronto Working Papers in Linguistics 10. 65-82. Sagey, Elizabeth (1986). The Representation of Features and Relations in Non-linear Phonology. Ph.D. dissertation, Massachusetts Institute of Technology. Tadmor, Uri and Yanti (2005). Complex oral-nasals as boundary markers in Traditional Jambi Malay. Presented at ISMIL 9. Maninjau, 27-29 July 2005 Walker, R. (1998). Nasalization, neutral segments, and opacity effects. PhD dissertation. UCSC. Weijer, Jerome M. van de & Frans Hinskens (2004). Segmental markedness and complexity - A cross-linguistic study of consonantal inventories . Presented at GLOW 2004.

Appendix 1: On the nature of the nasal/oral alternation
Partially-nasal stops become simple nasal/completely nasal stops when adjacent to a tauto-syllabic nasal(ised) vowel. (21) Jambi Malay - from (6) a. After oral vowels it(t)abm

‘black’

b. After nasalised vowels minũm

‘drink’


But, (22)

Guaraní - Walker (1998) a. /ro + mbo + γwatá/ ––––––> [rõmbo γwatá] ‘I made you walk’ I-you+CAUS+walk b. /ro+mbo+hendu/ ––––––> [rõmõhẽndu] ‘I made you hear’ I-you+CAUS+hear


This is true of all the (truly) partially-nasal stops.

14

Karthik Durvasula


Previous analyses have explained this by: •

Assuming a rule of [nasal] fusion within syllables (Piggott (1992)).

•

Assuming the relevant feature (SV) attaches to the syllable node, so naturally extends to all the linked segments. (Piggott and Hulst (1997)).


However, Piggott and Hulst (1997) reject (a) as the ‘rule’ does not capture the universal nature of the phenomenon.
Furthermore, Nasukawa (2005) rejects Piggott and Hulst (1997) based on a host of reasons.
Even a gestural overlap account of partially-nasal stops (a la Browman and Goldstein (1989)), whereby, the oral quality heard is a result of the overlap of the oral gesture of the vowel with a constriction in the oral cavity (of the next consonant), is unlikely. There are dialects of Jambi Malay, where the oral occlusion is NOT between the vowel and the following nasal consonant, yet they show the same basic nasal/oral alternation. (23) Alternation between nasals and post-stopped nasals in Dusun Dalam (Anderbeck 2003)11 p

a. nãm ‘six’

d. garam ‘salt’

b. diŋĩn

e. taon

‘cold’

c. barenãŋ ‘swim’

t

‘year’ k

f. jantuŋ ‘heart’


So, we are left with no explanation!!

11

This dialect (Dusun Dalam – DD) has post-occluded nasals essential in environments where the Tanjung Raden (TR) dialect has pre-occlusion. (i) DD TJ Meaning d d jalan jala n ‘road’ makant makadn ‘eat’ p p garam gara m ‘salt’ However, the phenomenon is not as consistent in Dusun Dalam (at least according to Anderbeck (2003)) as it is in Tanjung Raden.

Segmental Inventories, GLOW Workshop, 11th April 2007.

15


Within the present framework, one possible answer is in how the phonetic component performs gestural completions. •

Perhaps only a single gesture completes identically specified adjacent dimensions within a syllable as in (24a), and identically specified non-adjacent dimensions receive an equivalent (with respect to the identically specified dimensions) number of gestural completions as in (24b).

(24)a. Identically specified adjacent dimensions within a syllable. [syl X X ] ––––––––> | | SP SP

[syl X X ] | | SP SP \/ [nasal] b. Identically specified non-adjacent dimensions within a syllable. [syl X X X ] | | SP SP

––––––––>

[syl X X X ] | | SP SP | | [nasal/oral/Ø] [nasal/oral/Ø] (The actual completion will depend on the nature of stop contrast.)
What can this analysis buy us? •

Maybe, we can explain aspiration in sC cluster in English the same way.

(25) Gestural completion of sC clusters [syl X X … ] | | GW GW

––––––––>

[syl X X … ] | | GW GW \/ [spread glottis]

Appendix 2: Mixtec – a contradiction?
Mixtec is an Otomanguean language spoken in Mexico.
Dialects of Mixtec, at first sight, appear to contradict the analysis proposed in this paper.

16

Karthik Durvasula


Coatzospan Mixtec (spoken in the village of San Juan Coatzospan, Mexico) has a 3way stop distinction between simple (voiceless) stops, simple nasal stops and partially-nasal stops (Gerfen (1996)). (26)

a. ndĩĩ

‘you (familiar) will go down’

b. tĩĩ

‘sweat’

c. nĩnũ

‘up’


This should not be possible given our analysis as partially-nasal stops and the simple stops are argued to have the same representation.
However, there is no such problem – Iverson and Salmons (1996), based on phonological and phonetic evidence, independently come to the conclusion that the partially-nasal stop series is best analysed as a ‘voiced’ stop series underlyingly that are ‘enhanced’ with pre-nasalisation – hence, they are a case of ‘(phonetically) prenasalised voiced stops’.
The present analysis reaches a similar conclusion - the ‘prenasalised stops’

in

Coatzospan Mixtec do not behave like regular partially-nasal stops, they do not crucially participate in the nasal/oral alternation adjacent to tauto-syllabic nasal(ised) vowels, so they are not phonologically nasal.