Steve Parker

Sound level protrusions as physical correlates of sonority

Abstract A long-standing controversy in the interface between phonetics and phonology involves the nature of sonority: does it even exist and, if so, what are its phonetic correlates, and how can this be empirically demonstrated? This paper seeks to help resolve this problem by providing physical evidence supporting the sonority hierarchy. This is accomplished by reporting the results of a rigorous experiment measuring sound levels of realizations of all phonemes in English, Spanish, and Quechua. The obtained intensity values yield an overall mean Spearmans correlation of .91 with the proposed sonority indices. Consequently, one possible way to informally characterize sonority is in terms of a linear regression equation based on the observed intensity results. In light of these findings, the frequent claim that sonority lacks a reliable phonetic basis can no longer be maintained.

Non-optimal onsets in Chamicuro: an inventory maximised in coda position

The widely attested onset/coda asymmetry involves a situation in which the inventory of phonemes in syllable-final position in a particular language is a subset of those which contrast in onsets. The inverse of this pattern has been claimed to never occur (Goldsmith 1990, Beckman 1998). However, this prediction is falsified by Chamicuro, a Peruvian language in which /h/ and /[glottal plosive]/ are systematically restricted to coda position. Since no permutation of all known constraints can account for this unusual distribution, a new constraint is necessary. I propose that we invoke H AVE P LACE and subcategorise it for onsets. This positional markedness filter permits placeless (laryngeal) consonants to surface in codas, but blocks them in onsets. A beneficial side-effect of this analysis is that it preserves the onset/coda asymmetry while allowing /[glottal plosive]/ and /h/ to be the only principled exceptions to it.

ON THE BEHAVIOR OF DEFINITE ARTICLES IN CHAMICURO

This report describes the unique behavior of two clitic particles in a moribund Amazonian language. In Chamicuro na and ka are basically definite articles, yet they contrast for tense (ka indicates a past action while na is used in present and future contexts). The prosodic encliticization of na and ka is completely predictable from general phonological constraints alone and thus does not depend on any syntactic factors. Consequently, the stranding of an article in a different clause from its head noun follows automatically from the posited constraint system without any need to stipulate additional formal devices.*

The psychological reality of sonority in English

Abstract A long-standing controversy in the interface between phonology and other disciplines involves the nature of sonority. This paper seeks to help resolve this issue by showing that the sonority hierarchy is psychologically real to native speakers of English. This is accomplished by reporting the results of a rigorous and in-depth psycholinguistic experiment involving a common process of playful reduplication. A list of 99 hypothetical rhyming pairs such as roshy-toshy was evaluated by 332 subjects. Their task was to judge which order sounds more natural, e.g., roshy-toshy or toshy-roshy. The data confirm the crucial importance of sonority in accounting for the observed results. Specifically, the unmarked (preferred) pattern is for the morpheme beginning with the more sonorous segment in each pair to occur in absolute word-initial position. A generalized version of the Syllable Contact Law is utilized in a brief formal analysis of this phenomenon.

Initial and medial geminate trills in Arop-Lokep

Arop-Lokep, an Austronesian language of Papua New Guinea, has a contrast between a trilled /r/ and a geminate /rr/ in both intervocalic and word-initial positions. This paper presents the results of acoustic and perceptual experiments involving these segments. Specifically, we probe the statistical characteristics of the /r/ vs. /rr/ contrast across prosodic positions from the perspective of both the speaker and the listener. We measure trills in 24 different Arop-Lokep words, so the resulting data are quite robust and give an excellent idea of the variation inherent in the language. A novel finding is that both lengths of trill consist of more apical contacts word-initially than intervocalically. The scope of this work is unprecedented in that no previous instrumental study has examined both initial and medial geminate trills in the same language.

The acoustic qualities of Bora vowels.

Bora contains the vowel system /i e a o i w/, where /i/ is high, central, open, unrounded and /w/ is high, back, open, unrounded. A contrast between /i/ and /u/ is exceedingly rare in the languages of the world. An opposition between a central and a back vowel which are otherwise identical is theoretically significant since it shows that the binary feature [+/- back] is too weak to encode all phonological contrasts along the front/back dimension. However, previous analyses of languages containing such oppositions are based primarily on impressionistic transcriptions without instrumental confirmation. This study contributes to the debate by providing acoustic evidence supporting the minimal contrast between /i/ and /w/ in Bora. The results indicate that these phonemes have identical F1 values, but /i/ yields higher F2 and F3 measurements than /w/, in keeping with their reported difference in backness.

Chamicuro laterals: A case for Radical Underspecification

Abstract Two distinct phonological processes in Chamicuro converge on /i/ as being the unmarked, default vowel in this language: Stray Epenthesis inserts a word-final [i] to correct defective syllable structure, and an allophonic affrication of /1/ is triggered by all vowels except /i/. In the formalization of the latter rule, there is no simple and natural way to capture the environmental generalization given the conditions which the theory of Contrastive Underspecification places on the underlying vowel system. In terms of Radical Underspecification, however, the process can be formalized in a much more natural and straightforward way, thus providing additional confirmation for the empirical adequacy of the latter model.

A comparison of lip positions for /ɨ/ and /Ɯ/ in Bora

Bora is a Witotoan language spoken by about 750 persons in the Amazon jungle of Peru and 100 in Colombia. Its phonemic vowels are /i e a o ɨ Ɯ/ (Thiesen and Weber 2012). A contrast between a central and a back vowel which are otherwise identical is theoretically significant since it shows that a binary feature [ + /-back] is too weak to encode all phonological contrasts along the front/back dimension. The three high vowels of Bora have been acoustically confirmed with measurements of F1-F3 (Parker 2001), but there has been no articulatory investigation of these vowels. Impressionistically all of the Bora vowels except /o/ are articulated with unrounded lips. However, Ladefoged and Maddieson (1996) note that high back unrounded vowels in languages such as Japanese involve a gesture of lip compression or inrounding. Consequently, an important research question is whether the distinction between /ɨ/ and /Ɯ/ in Bora can be relegated to a difference in lip positions rather than to a primary contrast in tongue ...

Talking to Pets in Arara

In this article, we describe the phonological behavior of a unique series of fourteen language games in Arara, a Cariban language of Brazil. These are known mainly by elderly speakers only. The primary sociolinguistic function of the language games is to express solidarity and friendship with different pet animals, including dogs, birds, and various kinds of monkeys. Each language game is used when talking to or with a particular species of animal (or a closely related class of animals), and is characterized by a different formative (morpheme) added to the normal Arara base word. The formatives include prefixes, infixes, and autosegmentalized features such as nasalization and murmuring; some of these are not attested in the language other than in this context. We provide an informal analysis of the morphophonemic processes exhibited by a wide range of Arara forms to which each of the fourteen language games is applied in turn. We thereby document an interesting yet endangered aspect of the linguistic interaction between one specific group of indigenous people and their (potentially) nondomesticated pets.

The transparency and density of phonological rule interactions

Abstract This article introduces two new mathematical formulas designed to describe, compare, and contrast languages in terms of the overall markedness and complexity of their phonological rule components. The Transparency Index evaluates the four classical types of ordering relationships which obtain between individual pairs of rules, building on Kiparskys (Linguistic universals and linguistic change, Holt, Rinehart & Winston, 1968, Historical linguistics, University of Maryland Linguistics Program, 1971) classification of these as either feeding, bleeding, counterfeeding, or counterbleeding. Among these, feeding and bleeding orders are transparent in that they result in surface-true generalizations and are therefore natural and preferred, whereas counterbleeding and counterfeeding relationships lead to opaque (non-transparent) phonetic outcomes since the final surface representation appears to be a counterexample to one of the rules. For each language that we wish to analyze, the Transparency Index calculates the attested frequency of both types of transparent interactions grouped together (feeding and bleeding), and divides this sum by the total number of tokens of all four rule ordering types combined. The Density Index, on the other hand, computes the total number of crucial rule orderings (regardless of which kind) among each exhaustive, language-specific inventory of processes. These two statistical proposals are applied to a sample of seven genetically and areally diverse languages to illustrate their usefulness and rank each language along a continuum of relative typological markedness and simplicity. This study constitutes the first rigorous attempt to quantify the occurrence of different types of phonological rule relationships – including their degree of interaction and inherent rate of opacity – both within and across languages. It thus represents a step forward in extending an old line of inquiry into a topic which has recently spawned a renewed wave of interest, partly due to the problems it raises for standard Optimality Theory.