A field-based ultrasound and acoustic study of Iwaidja, an endangered Australian Aboriginal langu... more A field-based ultrasound and acoustic study of Iwaidja, an endangered Australian Aboriginal language, investigates the phonetic identity of nonnasal velar consonants in intervocalic position, where past work has proposed a [+continuant] vs. [−continuant] phonemic contrast. We analyze the putative contrast within a continuous phonetic space, defined by both acoustic and articulatory parameters, and find gradient variation: from more consonantal realizations, such as [ɰ], to more vocalic realizations, such as [a]. The distribution of realizations across lexical items and speakers does not support the proposed phonemic contrast. This case illustrates how lenition that is both phonetically gradient and variable across speakers and words can give the illusion of a contextually restricted phonemic contrast.*
Descriptions of Australian Aboriginal English list the neutralisation of the Standard English con... more Descriptions of Australian Aboriginal English list the neutralisation of the Standard English contrast between so-called voiced and voiceless stops as one characteristic feature. This paper reports on the results of an acoustic analysis of data collected in a production task by monolingual speakers of Standard Australian English in Sydney, of Aboriginal English on Croker Island, Northern Territory, and bilingual speakers of Iwaidja/Aboriginal English and Kunwinjku/Aboriginal English on Croker Island. The results show that average values for Voice Onset Time, the main correlate of the “stop voicing contrast” in English, and Closure Duration collected from Aboriginal speakers of English do not significantly differ from that of speakers of Standard Australian English, irrespective of language background. This result proves that the stop contrast is not neutralised by these Aboriginal speakers of English. However, it can be shown that phonetic voicing manifesting itself in Voice Termina...
One of the main characteristics of human languages is that they are subject to fundamental change... more One of the main characteristics of human languages is that they are subject to fundamental changes over time. However, because of the long transitional periods involved, the internal dynamics of such changes are typically inaccessible. Here, we present a new approach to examining language change via its connection to language comprehension. By means of an EEG experiment on Icelandic, a prominent current example of a language in transition, we show that the neurophysiological responses of native speakers already reflect projected changes that are not yet apparent in their overt behavior. Neurocognitive measures thus offer a means of predicting, rather than only retracing, language change.
This paper examines cross-linguistic influence in morphology among adult monolingual and heritage... more This paper examines cross-linguistic influence in morphology among adult monolingual and heritage speakers (Arabic-English and Chinese-English). Participants performed a task requiring them to form past tenses for English nonce words. Arabic-English bilinguals produced significantly more vowel change past tenses than either English monolinguals or Chinese-English bilinguals. We attribute the preponderance of vowel change past tenses to cross-linguistic influence of Arabic, as vowel change is a dominant morphological property in Arabic but not in English or Chinese. These results support dynamic models of bilingualism with constantly active and interacting languages and contribute to the phenomenology of crosslinguistic interference.
This paper aims to evaluate the Proto-Iwaidjan hypothesis, which proposes that several languages ... more This paper aims to evaluate the Proto-Iwaidjan hypothesis, which proposes that several languages in northwestern Arnhem Land in Northern Australia are genetically related. The evaluation is based on a rigorous application of the Comparative Method to provide an initial reconstruction of the Proto-Iwaidjan segmental inventory. We show that Amurdak, Iwaidja and Mawng are demonstrably genetically related. Given that Ilgar and Garig are very close in terms of grammar and vocabulary to Iwaidja, their membership of the Iwaidjan family is also supported. However, there is insufficient evidence to support the inclusion of Marrku, Wurrugu and Manangkari within the Iwaidjan language family.
Evaluation of hypotheses on genetic relationships depends on two factors: database size and crite... more Evaluation of hypotheses on genetic relationships depends on two factors: database size and criteria on correspondence quality. For hypotheses on remote relationships, databases are often small. Therefore, detailed consideration of criteria on correspondence quality is important. Hypotheses on remote relationships commonly involve greater geographical and temporal ranges. Consequently, we propose that there are two factors which are likely to play a greater role in comparing hypotheses of chance, contact and inheritance for remote relationships: (i) spatial distribution of corresponding forms; and (ii) language specific unpredictability in related paradigms. Concentrated spatial distributions disfavour hypotheses of chance, and discontinuous distributions disfavour contact hypotheses, whereas hypotheses of inheritance may accommodate both. Higher levels of language-specific unpredictability favour remote over recent transmission. We consider a remote relationship hypothesis, the Pro...
A book review of A Historical Phonology of English. Edinburgh Textbooks on the English Language –... more A book review of A Historical Phonology of English. Edinburgh Textbooks on the English Language – Advanced
As clearly stated in the introduction of this book, it is an introductory textbook aimed at stude... more As clearly stated in the introduction of this book, it is an introductory textbook aimed at students of German, specifically at students who are not set out to be linguists. It wants to get across the essentials, the basics of the sound structure of German (p. 7). However, the introduction also mentions that this book is suited for linguists and also for university teachers, as the presentation of the content “weicht…in einigen Punkten von den üblichen Darstellungen ab” [deviates at several points from more usual presentations]. Book review. The aim of this book is to show that “die Lautstrukturen des Deutschen einfacher sind, als sie üblicherweise gesehen und dargestellt werden” [the phonological structures of German are simpler as they are usually seen and portrayed]. I show that both of these statements make this book unique and almost revolutionary among modern textbooks
This paper is structured as follows. Section two discusses the question when the investigation of... more This paper is structured as follows. Section two discusses the question when the investigation of a contact etymology is warranted by the data and research history. Section 3 introduces the so-called Blueprint Principle, which is used to reconstruct possible contact languages in situation in which the contact language is unknown. In section 4 the core elements of a contact explanation are discussed, and section 5 outlines a methodological framework for the position of contact etymologies. The main points of this paper are summarized in section 6
A pivotal process in the loss of phonological quantity in West Germanic languages is what is trad... more A pivotal process in the loss of phonological quantity in West Germanic languages is what is traditionally known as Open Syllable Lengthening. Existing accounts have found no explanation for why languages such as English apply this change in less than 50% of the relevant cases. This paper presents the results of a corpus investigation of four West Germanic languages showing that whether Open Syllable Lengthening occurs in more than 50% of predicted cases correlates with the ratio of closed syllables with short vowels to open syllables with long vowels. We interpret this as the result of frequency effects that have markedly shaped the application of Open Syllable Lengthening in West Germanic. This has implications for phonological change in general, as well as for the relationship between stress and syllable structure in West Germanic languages.*
The aim of this paper is to integrate newly discovered modal categories and their respective para... more The aim of this paper is to integrate newly discovered modal categories and their respective paradigms in Amurdak, a non-Pama-Nyungan language belonging to the Iwaidjan language family, within the verb system of Amurdak and with a partial historical reconstruction of the verb system of Proto-Iwaidjan. The key challenge is to account for the heavy innovation of Amurdak in the area of modal categories, on the foundations of inherited material. In a first pass at this problem this paper will first provide a morphological analysis of the new data and then link them comparatively to other Iwaidjan languages, pointing out anchor points as well as current problems. The main point the paper wants to make is that, although Amurdak displays striking innovations with respect to other Iwaidjan languages – especially some of the categories presented here -, there are substantial links, which put Amurdak closer to the core of the Iwaidjan language family than previously assumed. There are at least three major connections between language documentation and linguistic theory that this paper highlights
A field-based ultrasound and acoustic study of Iwaidja, an endangered Australian Aboriginal langu... more A field-based ultrasound and acoustic study of Iwaidja, an endangered Australian Aboriginal language, investigates the phonetic identity of nonnasal velar consonants in intervocalic position, where past work has proposed a [+continuant] vs. [−continuant] phonemic contrast. We analyze the putative contrast within a continuous phonetic space, defined by both acoustic and articulatory parameters, and find gradient variation: from more consonantal realizations, such as [ɰ], to more vocalic realizations, such as [a]. The distribution of realizations across lexical items and speakers does not support the proposed phonemic contrast. This case illustrates how lenition that is both phonetically gradient and variable across speakers and words can give the illusion of a contextually restricted phonemic contrast.*
Descriptions of Australian Aboriginal English list the neutralisation of the Standard English con... more Descriptions of Australian Aboriginal English list the neutralisation of the Standard English contrast between so-called voiced and voiceless stops as one characteristic feature. This paper reports on the results of an acoustic analysis of data collected in a production task by monolingual speakers of Standard Australian English in Sydney, of Aboriginal English on Croker Island, Northern Territory, and bilingual speakers of Iwaidja/Aboriginal English and Kunwinjku/Aboriginal English on Croker Island. The results show that average values for Voice Onset Time, the main correlate of the “stop voicing contrast” in English, and Closure Duration collected from Aboriginal speakers of English do not significantly differ from that of speakers of Standard Australian English, irrespective of language background. This result proves that the stop contrast is not neutralised by these Aboriginal speakers of English. However, it can be shown that phonetic voicing manifesting itself in Voice Termina...
One of the main characteristics of human languages is that they are subject to fundamental change... more One of the main characteristics of human languages is that they are subject to fundamental changes over time. However, because of the long transitional periods involved, the internal dynamics of such changes are typically inaccessible. Here, we present a new approach to examining language change via its connection to language comprehension. By means of an EEG experiment on Icelandic, a prominent current example of a language in transition, we show that the neurophysiological responses of native speakers already reflect projected changes that are not yet apparent in their overt behavior. Neurocognitive measures thus offer a means of predicting, rather than only retracing, language change.
This paper examines cross-linguistic influence in morphology among adult monolingual and heritage... more This paper examines cross-linguistic influence in morphology among adult monolingual and heritage speakers (Arabic-English and Chinese-English). Participants performed a task requiring them to form past tenses for English nonce words. Arabic-English bilinguals produced significantly more vowel change past tenses than either English monolinguals or Chinese-English bilinguals. We attribute the preponderance of vowel change past tenses to cross-linguistic influence of Arabic, as vowel change is a dominant morphological property in Arabic but not in English or Chinese. These results support dynamic models of bilingualism with constantly active and interacting languages and contribute to the phenomenology of crosslinguistic interference.
This paper aims to evaluate the Proto-Iwaidjan hypothesis, which proposes that several languages ... more This paper aims to evaluate the Proto-Iwaidjan hypothesis, which proposes that several languages in northwestern Arnhem Land in Northern Australia are genetically related. The evaluation is based on a rigorous application of the Comparative Method to provide an initial reconstruction of the Proto-Iwaidjan segmental inventory. We show that Amurdak, Iwaidja and Mawng are demonstrably genetically related. Given that Ilgar and Garig are very close in terms of grammar and vocabulary to Iwaidja, their membership of the Iwaidjan family is also supported. However, there is insufficient evidence to support the inclusion of Marrku, Wurrugu and Manangkari within the Iwaidjan language family.
Evaluation of hypotheses on genetic relationships depends on two factors: database size and crite... more Evaluation of hypotheses on genetic relationships depends on two factors: database size and criteria on correspondence quality. For hypotheses on remote relationships, databases are often small. Therefore, detailed consideration of criteria on correspondence quality is important. Hypotheses on remote relationships commonly involve greater geographical and temporal ranges. Consequently, we propose that there are two factors which are likely to play a greater role in comparing hypotheses of chance, contact and inheritance for remote relationships: (i) spatial distribution of corresponding forms; and (ii) language specific unpredictability in related paradigms. Concentrated spatial distributions disfavour hypotheses of chance, and discontinuous distributions disfavour contact hypotheses, whereas hypotheses of inheritance may accommodate both. Higher levels of language-specific unpredictability favour remote over recent transmission. We consider a remote relationship hypothesis, the Pro...
A book review of A Historical Phonology of English. Edinburgh Textbooks on the English Language –... more A book review of A Historical Phonology of English. Edinburgh Textbooks on the English Language – Advanced
As clearly stated in the introduction of this book, it is an introductory textbook aimed at stude... more As clearly stated in the introduction of this book, it is an introductory textbook aimed at students of German, specifically at students who are not set out to be linguists. It wants to get across the essentials, the basics of the sound structure of German (p. 7). However, the introduction also mentions that this book is suited for linguists and also for university teachers, as the presentation of the content “weicht…in einigen Punkten von den üblichen Darstellungen ab” [deviates at several points from more usual presentations]. Book review. The aim of this book is to show that “die Lautstrukturen des Deutschen einfacher sind, als sie üblicherweise gesehen und dargestellt werden” [the phonological structures of German are simpler as they are usually seen and portrayed]. I show that both of these statements make this book unique and almost revolutionary among modern textbooks
This paper is structured as follows. Section two discusses the question when the investigation of... more This paper is structured as follows. Section two discusses the question when the investigation of a contact etymology is warranted by the data and research history. Section 3 introduces the so-called Blueprint Principle, which is used to reconstruct possible contact languages in situation in which the contact language is unknown. In section 4 the core elements of a contact explanation are discussed, and section 5 outlines a methodological framework for the position of contact etymologies. The main points of this paper are summarized in section 6
A pivotal process in the loss of phonological quantity in West Germanic languages is what is trad... more A pivotal process in the loss of phonological quantity in West Germanic languages is what is traditionally known as Open Syllable Lengthening. Existing accounts have found no explanation for why languages such as English apply this change in less than 50% of the relevant cases. This paper presents the results of a corpus investigation of four West Germanic languages showing that whether Open Syllable Lengthening occurs in more than 50% of predicted cases correlates with the ratio of closed syllables with short vowels to open syllables with long vowels. We interpret this as the result of frequency effects that have markedly shaped the application of Open Syllable Lengthening in West Germanic. This has implications for phonological change in general, as well as for the relationship between stress and syllable structure in West Germanic languages.*
The aim of this paper is to integrate newly discovered modal categories and their respective para... more The aim of this paper is to integrate newly discovered modal categories and their respective paradigms in Amurdak, a non-Pama-Nyungan language belonging to the Iwaidjan language family, within the verb system of Amurdak and with a partial historical reconstruction of the verb system of Proto-Iwaidjan. The key challenge is to account for the heavy innovation of Amurdak in the area of modal categories, on the foundations of inherited material. In a first pass at this problem this paper will first provide a morphological analysis of the new data and then link them comparatively to other Iwaidjan languages, pointing out anchor points as well as current problems. The main point the paper wants to make is that, although Amurdak displays striking innovations with respect to other Iwaidjan languages – especially some of the categories presented here -, there are substantial links, which put Amurdak closer to the core of the Iwaidjan language family than previously assumed. There are at least three major connections between language documentation and linguistic theory that this paper highlights
Ghost phoneme? The velar approximant in Iwaida: evidence from ultrasound data
Australian languag... more Ghost phoneme? The velar approximant in Iwaida: evidence from ultrasound data
Australian languages usually have labial, palatal, and retroflex approximant phonemes. In addition, it has been proposed that Iwaidja, a Non-Pama-Nyungan language spoken in North-Western Arnhem Land, has a velar phoneme which has been analysed variably as either an approximant /ɰ/ (Evans 2009: 160) or a fricative /ɣ / (Evans 2000: 99). This phoneme has a limited distribution, occurring only between [+continuant] segments. Across Australian languages, velar approximants are common phonetic realizations of the velar stop phoneme in intervocalic position, where stops, particularly velar and labial stops, tend to undergo lenition. To ascertain the phonetic nature of the proposed /ɰ ~ ɣ/ phoneme in Iwaidja, and its relation to lenited stop realisations, we collected acoustic and ultrasound data on /ɰ ~ ɣ/ and the velar stop /g/. We also collected data on the palatal stop /ɟ/ and approximant /j/, as a comparison. Ultrasound images and synchronized audio were collected in a field setting on Croker Island in the Northern Territory, Australia. Four speakers (1 female) participated in the study. Target words containing /g, ɰ ~ ɣ, ɟ, j/ in intervocalic position were elicited using objects pictured on a computer monitor. Ultrasound and audio data were recorded while participants named the pictures in a standardised carrier phrase. Ultrasound recordings were made with a GE 8C-RS ultrasound probe held at a 90 degree angle to the jaw in the mid-sagittal plane with a lightweight probe holder (Derrick et al., 2015). The probe was connected to a GE Logiq-E (version 11) ultrasound machine. Video output from the ultrasound machine went through an Epiphan VGA2USB Pro frame grabber to a laptop computer, which used FFMPEG running an X.264 encoder to synchronize video captured at 60Hz with audio from a Sennheiser MKH 416 microphone. Preliminary analysis (see Figure 1) indicates a clear distinction between articulation of consonants previously analysed as stops (blue circles), and as approximants (red squares), at both palatal (left panel) and velar (right panel) places of articulation. The figure compares edgetracks (Li et al. 2005) of 6-8 tokens per contrast in the same [a_a] context. The origin of the plot is the posterior portion of the tongue. The stop [ɟ] (blue circles, left panel) differs from the approximant [j] (red squares, left panel) in being more front and slightly higher. The right panel shows the stop-approximant contrast at the velar place of articulation. Although the velar series is more variable than the palatal series, the velar stop is, on average, higher (~2mm) than the velar approximant. Acoustic data provides clear evidence of closure for palatal stops but not for velar stops. The height of the tongue for /ɰ ~ ɣ/ is similar to the vowel /u/ in our data. Although more analysis is required, preliminary results do not support a consonantal contrast: /g/ vs /ɰ ~ɣ/. Rather, at least between identical vowels, it appears that sequences previously analyzed phonemically, for example as /aɰa ~ aɣa/, are phonetically overlong vowels [aaa]. Figure 2 shows that the duration of sequences historically analyzed as /_aɰa_/ occurring as [_aaa_] in our data are nearly three times as long as [a] occurring in stressed position. In some cases, there is historical evidence that these overlong vowels derive from sequences with a velar stop, such as /aga/ (e.g. Iwaidja Nahawaj vs. Bininj Gun-Wok Nawagaj ‘subsection name). This suggests a kind of compensatory lengthening. Cross-linguistically, such losses of approximants in syllable onset position with simultaneous compensatory lengthening are rare (see e.g. Kümmel 2007: 126f) and cannot readily be explained by mora preservation (e.g. Hayes 1989) or some kind of a phonetic merger analysis (e.g. de Chene & Anderson 1979) Thus, the Iwaidja data is theoretically challenging not because of the presence of a velar approximant, but because the timing slot of onset velar stops appears to have been preserved over time even as the gestural structure has been reduced to the point of complete lenition.
Figure 1: Palatal and velar approximants (red) vs. stops (blue)
Figure: 2: Vowel length of VhV vs. V
References Clements, G.N. & S.J. Keyser. 1983. CV phonology. The MIT Press. de Chene, E. B. & S. R. Anderson. 1979. Compensatory Lengthening. Language 55:505- 535. Derrick, D., C. Best, R. Fiasson. (2015) Non-metallic ultrasound probe holder for co-collection and co-registration with EMA. Proceedings of ICPHS. Evans, N. (2000). Iwaidjan, a very un-Australian language family. Linguistic Typology, 4(2), 91-142. Evans, N. (2009). Doubled up all over again: borrowing, sound change and reduplication in Iwaidjan. Morphology 19, 159-176. Hayes, 8. 1989. Compensatory Lengthening in Moraic Phonology. Linguistic lnquiry, ‘20: 253-306 Li, M., Kambhamettu, C., & Stone, M. (2005). Automatic contour tracking in ultrasound images. Clinical linguistics & phonetics, 19(6-7), 545-554. Kümmel, M. J. 2007. Konsonantenwandel. Wiesbaden: Reichert
Stops in Croker Island English: making use of a multilingual repertoire
The phonetics of Aborigi... more Stops in Croker Island English: making use of a multilingual repertoire
The phonetics of Aboriginal English is still under-researched (Eades 2014: 438). Apart from general overviews (see e.g. Butcher 2008; Fletcher & Butcher 2014), studies in which Kriol is the focus (Jones & Meakins 2013) and studies on Kriol only (Baker et al. 2014), instrumental phonetic research on varieties of Aboriginal English appears to be lacking. However, studying the phonetic detail of Aboriginal English is of great significance for mapping out the sociolinguistic variation of English, especially in relation to the linguistic contribution of Indigenous languages. Furthermore, it is relevant to a phenomenology of contact-induced change, due to the synchronic and diachronic complexity of the contact situation. For instance it has been speculated that the influence of Indigenous languages may be the reason that stops in some varieties of Kriol have a longer duration than in Standard Australian English (Baker et al. 2014: 328). Though suggestive, this remains an unproven assumption (see Jones & Meakins 2013: 216-17). Studies of contact-induced change in the area of phonetics and phonology have focused on transfer and convergence effects (e.g. Antoniou et al. 2011), rather than bilingual correspondence-matching and optimisation strategies, despite clear indication for such phenomena (Muysken 2013: 725). This study investigates two phonetic parameters, Voice Onset Time (VOT) and Constriction Duration (CD), in Aboriginal English, Iwaidja and Kunwinjku, in order to answer the following research questions: 1. Does English spoken by Iwaidja and Kunwinjku speakers differ from general non-Aboriginal English and if so how can differences be described and explained? 2. Does English spoken by Iwaidja speakers differ from English spoken by Kunwinjku speakers with respect to these two parameters (see e.g. Antoniou et al. 2011: on the transferability of VOT settings), and if so, how can this be described and explained? Acoustic data were collected on Croker Island, NT. Eight speakers (2 female, 2 male Iwaidja-English bilinguals, 2 female, 2 male Kunwinjku bilinguals) participated in the study. Suitable target words were elicited in English, Iwaidja and Kunwinjku either by written stimuli or by shadowing with stop phonemes in initial, medial and final position embedded in a natural carrier phrase controlling for the phonological environment (target words per condition). Recordings were made with a Countryman EMW microphone using an iPad with iRigPro preamp with a 16-bit sampling depth and a 48kHz sampling rate. Preliminary results suggest both Iwaidja English and Kunwinjku English make use of CD as well as VOT to mark the phonemic voicing contrast in medial position, although in an asymmetrical way (see Figures 1 and 2): while in both varieties VOT and CD for voiced stops roughly pattern with non-Aboriginal English, CD in voiceless stops is significantly longer than in non-Aboriginal English. VOT and CD of Iwaidja and Kunwinjku English voiced stops matches with the relevant values of stops in Iwaidja and singleton stops in Kunwinjku. The use of a longer CD for voiceless stops in Kunwinjku English can be understood against the background of Kunwinjku: speakers seem to map the main phonetic correlate of the Kunwinjku long vs. short stops onto the voiceless vs. voiced stop contrast in English and use that in addition to VOT. However, this is not a likely explanation for Iwaidja English, as Iwaidja only has a single stop series. In fact, it would be expected that Iwaidja English neutralises the voicing distinction (Butcher 2008: 627). Instead, Iwaidja English differentiates “voiced” and “voiceless” stops also by CD and VOT. We interpret the Kunwinjku and Iwaidja English data as an “optimisation” or enhancement strategy in the sense of Muysken (2013): speakers capitalise on correspondences in all available languages and use them simultaneously. Given that the English voiced stops match the Iwaidja single stop series in terms of CD, we suggest that the phonetically slightly longer VOT of English voiceless stops is exaggerated to provide a sufficient contrast. Thus, this paper shows that convergence and transfer are inadequate in explaining the CD and VOT measured in Iwaidja and Kunwinjku English. Instead, language-specific strategies are used to find correspondences for phonological contrasts within the languages available.
Figure 1: Medial position constriction duration in Iwaidja English, Kunwinjku English, Iwaidja and Kunwinjku (IPA symbols refer to phonetic features rather than phonemes)
Figure 2: Medial position VOT in Iwaidja English, Kunwinjku English, Iwaidja and Kunwinjku (IPA symbols refer to phonetic features rather than phonemes)
References Antoniou, Mark, Catherine T. Best, Michael D. Tyler & Christian Kroos. 2011. Inter-language interference in VOT production by L2-dominant bilinguals: Asymmetries in phonetic code-switching. Journal of Phonetics 38.558-70. Baker, Brett, Rikke Bundgaard-Nielsen & Simone Graetzer. 2014. The Obstruent Inventory of Roper Kriol. Australian Journal of Linguistics 34.307-44. Butcher, Andrew. 2008. Linguistic aspects of Australian Aboriginal English. Clinical Linguistics & Phonetics 22.625-42. Eades, Diana. 2014. Aboriginal English. The languages and linguistics of Australia: a comprehensive guide, ed. by H. Koch & R. Nordlinger, 417-47. Berlin/New York: De Gruyter Mouton. Fletcher, Janet & Andrew Butcher. 2014. Sound patterns of Australian Languages. The languages and linguistics of Australia: a comprehensive guide, ed. by H. Koch & R. Nordlinger. Berlin/New York: De Gruyter Mouton. Jones, C. & F. Meakins. 2013. Australian Journal of Linguistics 33.196. Muysken, Pieter. 2013. Language contact outcomes as the result of bilingual optimization strategies. Bilingualism: Language and Cognition 16.709-30.
Usage-based and other functional theories of language change and especially research on grammatic... more Usage-based and other functional theories of language change and especially research on grammaticalisation generally view frequency as the central factor in the emergence of new linguistic structure. What is, however, less clear is what motivates changes in frequency in the first place. Notions such as iconicity (Haiman 1983), relevance and generality (Bybee 1985), prototypicality (Bybee 2010), and system optimisation (Haspelmath 2014) have remained vague and difficult to operationalise in a testable model. In our paper, we propose that selectional restrictions on the combinability of words may provide a better foundation for such a model. Specifically, we develop a predictive model of how selectional restrictions affect frequency, and thus language change, and apply it in a case study. There are various kinds of “selectional” restrictions that operate below the level of syntax and that influence what can or cannot be combined (see e.g. Asher 2011, Jacobs 2006 for the theoretical modelling of selectional restrictions). One common type are restrictions imposed by semantics, especially tense and aspect, such as the well-known fact that present tense and perfective viewpoints generally lead to marked readings if they are possible at all (Smith 1997). Our hypothesis is that such selectional restrictions may be a factor behind changes in frequency, as they funnel what can be possible candidates for chunking and development into prefabs/constructions. Since some combinations will either be outright impossible and others will be less frequent due to their marked nature, they may form less optimal candidates. We hypothesise that mapping out selectional restrictions can give an insight into what may be considered frequent as opposed to infrequent constructions, and why this might be the case. We tested this hypothesis in a study on the development of the passive construction in German and English. Our investigation focused on the early Old English and Old High German periods (850-950 CE) and the end of the Old English period and the beginning of the Middle High German period (1050-1150 CE). Previous work established that both languages start out with a similar inventory of fully compositional constructions of a copula verb plus a past participle, whose readings depend on their aspectuo-temporal configurations (Mailhammer & Smirnova 2013). However, the constructions in the two languages subsequently diverge in promoting different copula verbs to auxiliary verbs in passive diatheses. Starting from the assumption that constructions evolve from the range of possible combinations, we investigated the available configurations of copula constructions with respect to tense and aspect parameters. The aim was first to see what kinds of configurations might be more prototypical for a copula construction with passive reading and how other readings were distributed in the corpus and whether this could provide any clues as to their further development in both languages. The model we developed to describe these constructions computes the compositional meaning as a function of the semantic properties of its elements, especially tense and aktionsart (similarly to Randall & Jones 2014). Given the prototypical function of a copula construction all state readings are prototypical representatives, but precisely those are unlikely candidates to develop into passive diatheses, as passives typically denote processes. Thus, the marked reading obtained in configurations with a process outcome, is expected to increase if the end-result is to be a passive diathesis. This means that the prediction is that typically only combination with ‘become’ as the copula verb (especially in the present tense) and past participles of transitive activity or accomplishment verbs are the first to increase in frequency and to develop into passives but not combinations with ‘be’, ‘become’, or stative and achievement verbs. Applied to German and English, this predicts that the above copula constructions should be distributed as follows:
(1) Distributional predictions of type of copula and aktionsart of participle a. German: initially low frequency, and then increase in frequency of combinations with ‘become’ in the present tense as well as past participles of transitive activity and accomplishment verbs b. English: initially low frequency, and no increase in later periods; perhaps even an increase in resultative-stative combinations
The corpus data confirm our hypotheses: Synchronically observable restrictions in combinability of copula constructions with passive readings in Old English and Old High German leave two developmental paths: maintenance of a copula construction and grammaticalisation into a passive. The predictions with respect to the configuration that permits the second development, namely combinations of present tense ‘become’ with activity verbs and semelfactives is borne out for German, which eventually grammaticalises ‘become’ into a passive auxiliary, while the ‘be’ construction remains a copula construction. For English the prediction is not borne out directly: the fact that OE weorðan remains a copula construction is predicted from the distribution at the end of the OE period, but that the b-forms of bēon do not grammaticalise into a passive has reasons outside the aspectuo-temporal restrictions imposed on the construction. The implications of this case study are that selectional restrictions are indeed shapers of language change and that hypotheses built on modelling the compositional values of lexical semantics can make testable predictions as to the paths of structural development.
References Asher, Nicholas. 2011. Lexial Meaning in Context. Cambridge: Cambridge University Press. Bybee, Joan. 1985. Morphology: A Study of the Relation between Meaning and Form. Amsterdam/Philadelphia: John Benjamins. Bybee, Joan. 2010. Language, Usage and Cognition. Cambridge: Cambridge University Press. Haiman, John. 1983. Iconic and economic motivation. Language 59, 781-819. Haspelmath, Martin. 2014. On system pressure competing with economic motivation. In Brian MacWhinney, Andrej L. Malchukov & Edith A. Moravcsik (eds.), Competing motivations in grammar and usage, 197–208. Oxford: Oxford University Press. Jacobs, Joachim. 2006. Die Problematik der Valenzebenen. In Hans-Werner Eroms (ed.), Dependenz und Valenz, 2 vols, vol. 2 (Handbücher zur Sprach- und Kommunikationswissenschaft 25), 378-99. Berlin/New York: De Gruyter. Mailhammer, Robert & Elena Smirnova. 2013. Incipient Grammaticalisation: sources of passive constructions in English and German. In Gabriele Diewald, Ilse Wischer & Leena Kahlas-Tarkka (eds.), Comparative Studies in Early Germanic Languages: with a Focus on Verbal Categories, 41-69. Amsterdam/Philadelphia: John Benjamins. Randall, William & Howard Jones. 2014. On the early origins of the Germanic preterite presents. Transactions of the Philological Society 0, 1-40. Smith, Carlota. 1997. The Parameter of Aspect, 2nd edn. Dordrecht: Kluwer.
Previous work has unearthed much about how grammaticalisation proceeds, but little to how it star... more Previous work has unearthed much about how grammaticalisation proceeds, but little to how it starts. This paper focuses on the development of passive auxiliaries from copula verbs, offering a model of the initial steps in the grammaticalisation process towards passive. The model emphasises the role of semantic factors in the compositional structure of the relevant copula constructions. We present evidence and a theoretical foundation for why atypical – " misfit " – constellations of constructions may begin the grammaticalisation path by prompting reanalysis. The data come from Old and Middle English and Old and Middle High German. We investigated the aktionsart combinations of copula constructions with 'become' (OHG werdan, OE weorðan) from Old English and Old High German to Middle English and Middle High German. The crucial difference between English and German is that combinations of 'become' and activity/semelfactive verbs become more frequent in German, which is what our theoretical model predicts.
Uploads
Papers by Robert Mailhammer
Australian languages usually have labial, palatal, and retroflex approximant phonemes. In addition, it has been proposed that Iwaidja, a Non-Pama-Nyungan language spoken in North-Western Arnhem Land, has a velar phoneme which has been analysed variably as either an approximant /ɰ/ (Evans 2009: 160) or a fricative /ɣ / (Evans 2000: 99). This phoneme has a limited distribution, occurring only between [+continuant] segments. Across Australian languages, velar approximants are common phonetic realizations of the velar stop phoneme in intervocalic position, where stops, particularly velar and labial stops, tend to undergo lenition. To ascertain the phonetic nature of the proposed /ɰ ~ ɣ/ phoneme in Iwaidja, and its relation to lenited stop realisations, we collected acoustic and ultrasound data on /ɰ ~ ɣ/ and the velar stop /g/. We also collected data on the palatal stop /ɟ/ and approximant /j/, as a comparison.
Ultrasound images and synchronized audio were collected in a field setting on Croker Island in the Northern Territory, Australia. Four speakers (1 female) participated in the study. Target words containing /g, ɰ ~ ɣ, ɟ, j/ in intervocalic position were elicited using objects pictured on a computer monitor. Ultrasound and audio data were recorded while participants named the pictures in a standardised carrier phrase. Ultrasound recordings were made with a GE 8C-RS ultrasound probe held at a 90 degree angle to the jaw in the mid-sagittal plane with a lightweight probe holder (Derrick et al., 2015). The probe was connected to a GE Logiq-E (version 11) ultrasound machine. Video output from the ultrasound machine went through an Epiphan VGA2USB Pro frame grabber to a laptop computer, which used FFMPEG running an X.264 encoder to synchronize video captured at 60Hz with audio from a Sennheiser MKH 416 microphone.
Preliminary analysis (see Figure 1) indicates a clear distinction between articulation of consonants previously analysed as stops (blue circles), and as approximants (red squares), at both palatal (left panel) and velar (right panel) places of articulation. The figure compares edgetracks (Li et al. 2005) of 6-8 tokens per contrast in the same [a_a] context. The origin of the plot is the posterior portion of the tongue. The stop [ɟ] (blue circles, left panel) differs from the approximant [j] (red squares, left panel) in being more front and slightly higher. The right panel shows the stop-approximant contrast at the velar place of articulation. Although the velar series is more variable than the palatal series, the velar stop is, on average, higher (~2mm) than the velar approximant.
Acoustic data provides clear evidence of closure for palatal stops but not for velar stops. The height of the tongue for /ɰ ~ ɣ/ is similar to the vowel /u/ in our data. Although more analysis is required, preliminary results do not support a consonantal contrast: /g/ vs /ɰ ~ɣ/. Rather, at least between identical vowels, it appears that sequences previously analyzed phonemically, for example as /aɰa ~ aɣa/, are phonetically overlong vowels [aaa]. Figure 2 shows that the duration of sequences historically analyzed as /_aɰa_/ occurring as [_aaa_] in our data are nearly three times as long as [a] occurring in stressed position.
In some cases, there is historical evidence that these overlong vowels derive from sequences with a velar stop, such as /aga/ (e.g. Iwaidja Nahawaj vs. Bininj Gun-Wok Nawagaj ‘subsection name). This suggests a kind of compensatory lengthening. Cross-linguistically, such losses of approximants in syllable onset position with simultaneous compensatory lengthening are rare (see e.g. Kümmel 2007: 126f) and cannot readily be explained by mora preservation (e.g. Hayes 1989) or some kind of a phonetic merger analysis (e.g. de Chene & Anderson 1979) Thus, the Iwaidja data is theoretically challenging not because of the presence of a velar approximant, but because the timing slot of onset velar stops appears to have been preserved over time even as the gestural structure has been reduced to the point of complete lenition.
Figure 1: Palatal and velar approximants (red) vs. stops (blue)
Figure: 2: Vowel length of VhV vs. V
References
Clements, G.N. & S.J. Keyser. 1983. CV phonology. The MIT Press.
de Chene, E. B. & S. R. Anderson. 1979. Compensatory Lengthening. Language 55:505-
535.
Derrick, D., C. Best, R. Fiasson. (2015) Non-metallic ultrasound probe holder for co-collection and co-registration with EMA. Proceedings of ICPHS.
Evans, N. (2000). Iwaidjan, a very un-Australian language family. Linguistic Typology, 4(2), 91-142.
Evans, N. (2009). Doubled up all over again: borrowing, sound change and reduplication in Iwaidjan. Morphology 19, 159-176.
Hayes, 8. 1989. Compensatory Lengthening in Moraic Phonology. Linguistic lnquiry,
‘20: 253-306
Li, M., Kambhamettu, C., & Stone, M. (2005). Automatic contour tracking in ultrasound images. Clinical linguistics & phonetics, 19(6-7), 545-554.
Kümmel, M. J. 2007. Konsonantenwandel. Wiesbaden: Reichert
The phonetics of Aboriginal English is still under-researched (Eades 2014: 438). Apart from general overviews (see e.g. Butcher 2008; Fletcher & Butcher 2014), studies in which Kriol is the focus (Jones & Meakins 2013) and studies on Kriol only (Baker et al. 2014), instrumental phonetic research on varieties of Aboriginal English appears to be lacking. However, studying the phonetic detail of Aboriginal English is of great significance for mapping out the sociolinguistic variation of English, especially in relation to the linguistic contribution of Indigenous languages. Furthermore, it is relevant to a phenomenology of contact-induced change, due to the synchronic and diachronic complexity of the contact situation. For instance it has been speculated that the influence of Indigenous languages may be the reason that stops in some varieties of Kriol have a longer duration than in Standard Australian English (Baker et al. 2014: 328). Though suggestive, this remains an unproven assumption (see Jones & Meakins 2013: 216-17). Studies of contact-induced change in the area of phonetics and phonology have focused on transfer and convergence effects (e.g. Antoniou et al. 2011), rather than bilingual correspondence-matching and optimisation strategies, despite clear indication for such phenomena (Muysken 2013: 725).
This study investigates two phonetic parameters, Voice Onset Time (VOT) and Constriction Duration (CD), in Aboriginal English, Iwaidja and Kunwinjku, in order to answer the following research questions:
1. Does English spoken by Iwaidja and Kunwinjku speakers differ from general non-Aboriginal English and if so how can differences be described and explained?
2. Does English spoken by Iwaidja speakers differ from English spoken by Kunwinjku speakers with respect to these two parameters (see e.g. Antoniou et al. 2011: on the transferability of VOT settings), and if so, how can this be described and explained?
Acoustic data were collected on Croker Island, NT. Eight speakers (2 female, 2 male Iwaidja-English bilinguals, 2 female, 2 male Kunwinjku bilinguals) participated in the study. Suitable target words were elicited in English, Iwaidja and Kunwinjku either by written stimuli or by shadowing with stop phonemes in initial, medial and final position embedded in a natural carrier phrase controlling for the phonological environment (target words per condition). Recordings were made with a Countryman EMW microphone using an iPad with iRigPro preamp with a 16-bit sampling depth and a 48kHz sampling rate.
Preliminary results suggest both Iwaidja English and Kunwinjku English make use of CD as well as VOT to mark the phonemic voicing contrast in medial position, although in an asymmetrical way (see Figures 1 and 2): while in both varieties VOT and CD for voiced stops roughly pattern with non-Aboriginal English, CD in voiceless stops is significantly longer than in non-Aboriginal English. VOT and CD of Iwaidja and Kunwinjku English voiced stops matches with the relevant values of stops in Iwaidja and singleton stops in Kunwinjku. The use of a longer CD for voiceless stops in Kunwinjku English can be understood against the background of Kunwinjku: speakers seem to map the main phonetic correlate of the Kunwinjku long vs. short stops onto the voiceless vs. voiced stop contrast in English and use that in addition to VOT. However, this is not a likely explanation for Iwaidja English, as Iwaidja only has a single stop series. In fact, it would be expected that Iwaidja English neutralises the voicing distinction (Butcher 2008: 627). Instead, Iwaidja English differentiates “voiced” and “voiceless” stops also by CD and VOT. We interpret the Kunwinjku and Iwaidja English data as an “optimisation” or enhancement strategy in the sense of Muysken (2013): speakers capitalise on correspondences in all available languages and use them simultaneously. Given that the English voiced stops match the Iwaidja single stop series in terms of CD, we suggest that the phonetically slightly longer VOT of English voiceless stops is exaggerated to provide a sufficient contrast.
Thus, this paper shows that convergence and transfer are inadequate in explaining the CD and VOT measured in Iwaidja and Kunwinjku English. Instead, language-specific strategies are used to find correspondences for phonological contrasts within the languages available.
Figure 1: Medial position constriction duration in Iwaidja English, Kunwinjku English, Iwaidja and Kunwinjku (IPA symbols refer to phonetic features rather than phonemes)
Figure 2: Medial position VOT in Iwaidja English, Kunwinjku English, Iwaidja and Kunwinjku (IPA symbols refer to phonetic features rather than phonemes)
References
Antoniou, Mark, Catherine T. Best, Michael D. Tyler & Christian Kroos. 2011. Inter-language interference in VOT production by L2-dominant bilinguals: Asymmetries in phonetic code-switching. Journal of Phonetics 38.558-70.
Baker, Brett, Rikke Bundgaard-Nielsen & Simone Graetzer. 2014. The Obstruent Inventory of Roper Kriol. Australian Journal of Linguistics 34.307-44.
Butcher, Andrew. 2008. Linguistic aspects of Australian Aboriginal English. Clinical Linguistics & Phonetics 22.625-42.
Eades, Diana. 2014. Aboriginal English. The languages and linguistics of Australia: a comprehensive guide, ed. by H. Koch & R. Nordlinger, 417-47. Berlin/New York: De Gruyter Mouton.
Fletcher, Janet & Andrew Butcher. 2014. Sound patterns of Australian Languages. The languages and linguistics of Australia: a comprehensive guide, ed. by H. Koch & R. Nordlinger. Berlin/New York: De Gruyter Mouton.
Jones, C. & F. Meakins. 2013. Australian Journal of Linguistics 33.196.
Muysken, Pieter. 2013. Language contact outcomes as the result of bilingual optimization strategies. Bilingualism: Language and Cognition 16.709-30.
There are various kinds of “selectional” restrictions that operate below the level of syntax and that influence what can or cannot be combined (see e.g. Asher 2011, Jacobs 2006 for the theoretical modelling of selectional restrictions). One common type are restrictions imposed by semantics, especially tense and aspect, such as the well-known fact that present tense and perfective viewpoints generally lead to marked readings if they are possible at all (Smith 1997). Our hypothesis is that such selectional restrictions may be a factor behind changes in frequency, as they funnel what can be possible candidates for chunking and development into prefabs/constructions. Since some combinations will either be outright impossible and others will be less frequent due to their marked nature, they may form less optimal candidates. We hypothesise that mapping out selectional restrictions can give an insight into what may be considered frequent as opposed to infrequent constructions, and why this might be the case.
We tested this hypothesis in a study on the development of the passive construction in German and English. Our investigation focused on the early Old English and Old High German periods (850-950 CE) and the end of the Old English period and the beginning of the Middle High German period (1050-1150 CE). Previous work established that both languages start out with a similar inventory of fully compositional constructions of a copula verb plus a past participle, whose readings depend on their aspectuo-temporal configurations (Mailhammer & Smirnova 2013). However, the constructions in the two languages subsequently diverge in promoting different copula verbs to auxiliary verbs in passive diatheses.
Starting from the assumption that constructions evolve from the range of possible combinations, we investigated the available configurations of copula constructions with respect to tense and aspect parameters. The aim was first to see what kinds of configurations might be more prototypical for a copula construction with passive reading and how other readings were distributed in the corpus and whether this could provide any clues as to their further development in both languages. The model we developed to describe these constructions computes the compositional meaning as a function of the semantic properties of its elements, especially tense and aktionsart (similarly to Randall & Jones 2014).
Given the prototypical function of a copula construction all state readings are prototypical representatives, but precisely those are unlikely candidates to develop into passive diatheses, as passives typically denote processes. Thus, the marked reading obtained in configurations with a process outcome, is expected to increase if the end-result is to be a passive diathesis. This means that the prediction is that typically only combination with ‘become’ as the copula verb (especially in the present tense) and past participles of transitive activity or accomplishment verbs are the first to increase in frequency and to develop into passives but not combinations with ‘be’, ‘become’, or stative and achievement verbs.
Applied to German and English, this predicts that the above copula constructions should be distributed as follows:
(1) Distributional predictions of type of copula and aktionsart of participle
a. German: initially low frequency, and then increase in frequency of combinations with ‘become’ in the present tense as well as past participles of transitive activity and accomplishment verbs
b. English: initially low frequency, and no increase in later periods; perhaps even an increase in resultative-stative combinations
The corpus data confirm our hypotheses: Synchronically observable restrictions in combinability of copula constructions with passive readings in Old English and Old High German leave two developmental paths: maintenance of a copula construction and grammaticalisation into a passive. The predictions with respect to the configuration that permits the second development, namely combinations of present tense ‘become’ with activity verbs and semelfactives is borne out for German, which eventually grammaticalises ‘become’ into a passive auxiliary, while the ‘be’ construction remains a copula construction. For English the prediction is not borne out directly: the fact that OE weorðan remains a copula construction is predicted from the distribution at the end of the OE period, but that the b-forms of bēon do not grammaticalise into a passive has reasons outside the aspectuo-temporal restrictions imposed on the construction.
The implications of this case study are that selectional restrictions are indeed shapers of language change and that hypotheses built on modelling the compositional values of lexical semantics can make testable predictions as to the paths of structural development.
References
Asher, Nicholas. 2011. Lexial Meaning in Context. Cambridge: Cambridge University Press.
Bybee, Joan. 1985. Morphology: A Study of the Relation between Meaning and Form. Amsterdam/Philadelphia: John Benjamins.
Bybee, Joan. 2010. Language, Usage and Cognition. Cambridge: Cambridge University Press.
Haiman, John. 1983. Iconic and economic motivation. Language 59, 781-819.
Haspelmath, Martin. 2014. On system pressure competing with economic motivation. In Brian MacWhinney, Andrej L. Malchukov & Edith A. Moravcsik (eds.), Competing motivations in grammar and usage, 197–208. Oxford: Oxford University Press.
Jacobs, Joachim. 2006. Die Problematik der Valenzebenen. In Hans-Werner Eroms (ed.), Dependenz und Valenz, 2 vols, vol. 2 (Handbücher zur Sprach- und Kommunikationswissenschaft 25), 378-99. Berlin/New York: De Gruyter.
Mailhammer, Robert & Elena Smirnova. 2013. Incipient Grammaticalisation: sources of passive constructions in English and German. In Gabriele Diewald, Ilse Wischer & Leena Kahlas-Tarkka (eds.), Comparative Studies in Early Germanic Languages: with a Focus on Verbal Categories, 41-69. Amsterdam/Philadelphia: John Benjamins.
Randall, William & Howard Jones. 2014. On the early origins of the Germanic preterite presents. Transactions of the Philological Society 0, 1-40.
Smith, Carlota. 1997. The Parameter of Aspect, 2nd edn. Dordrecht: Kluwer.