HistLing Seminar Series 1, ANU, 2004
Normal time: Fridays, 11.00am-12.30pm
Nadel Room (Seminar Room C), HC Coombs Building (no.9),
Australian National University
(Dept. Linguistics, RSPAS, ANU)
On the history of the Papuan-speaking populations of Near Oceania: What can linguistics add to the evidence of archaeology and other historical disciplines?
Papuan¹ is a useful label for the various families of non-Austronesian languages that are sandwiched between Australia to the south and the vast Austronesian-speaking area to the west, north and east. These almost certainly represent continuations of linguistic stocks that have been in Near Oceania (New Guinea and northwest Island Melanesia) for many millennia before Austronesian speakers entered this region some 3500-3300 years ago. Evidence concerning the prehistory of Near Oceania comes from a range of disciplines, including archaeology, palynology, historical linguistics, population genetics, paleozoology, geomorphology and human geography. The question we will address is: To what extent does linguistic evidence corroborate, contradict or add to the testimony of other disciplines concerning the history of the Papuan-speaking populations of Near Oceania? I will review recent work in historical linguistics and comment on how this squares with recent interpretations of evidence from other fields.
On Frequency and Finiteness: a universal path from pragmatics to syntax?
Despite some clear differences between Romanian and most other Romance languages in their use of the infinitive, remarkable similarities can be observed regarding the distribution of subordinate prepositional infinitives and their corresponding finite clauses: certain semantic notions, most clearly purpose, facilitate the use of the infinitive, whilst others, notably concessivity, show a strong resistance to infinitival subordination across the Romance languages. This is particularly interesting if one takes into account that this construction is not inherited from Latin and has emerged independently at different times in the languages' history.
In this paper it is argued that these distributional patterns develop as a function of the frequency of subject co-reference found with the respective semantic notions. A diachronic statistical analysis of Spanish and Romanian shows that the essentially extra-linguistically determined likelihood of co-reference with a given semantic notion remains virtually constant through time, but that a comparatively high frequency of subject co-reference goes hand in hand with an increasing preference for infinitival constructions. This clearly stems from the fact that the subject of prepositional infinitive clauses is, as a default, co-indexed with that of the main clause. But the link between semantic notion and finiteness develops its own dynamics by expanding beyond this default pattern. The fact that some of the resulting distribution patterns are cross-linguistically common raises the question whether this path, from pragmatically determined usage frequency to syntactic structure, might be a more wide-spread or even universal phenomenon.
Mood swings and downward migration in South Slavic and Ngumpin-Yapa (Pama-Nyungan)
The paper focuses on some similarities in parametric variation between Ngumpin-Yapa and Indo-European subgroups whose history is better attested, (Slavic, Romance). Ngumpin-Yapa has been well established as a sub-group by the standard comparative method (McConvell & Laughren 2004) and some preliminary work has been done towards reconstruction of the syntax of the proto-language (eg McConvell 1996). In Slavic, Romance and Ngumpin-Yapa pronominal clitics are widely found, but their placement varies between different languages and dialects. Here I am mainly concerned with the behaviour of imperative and other non-indicative clauses. As well as exploring the history of clitic placement in these clause types comparatively, I also look at how the Ng-Y contributes towards debate about whether Minimalist or Optimality Theoretic approaches are best suited to explaining the observed variation and the patterns of historical change in the syntax.
In Ngumpin-Yapa, imperatives in (a) Yapa (Warlpiri & Warlmanpa) require pronominal enclitic placement in clausal second position (2P) but (b) Gurindji and some other Ngumpin languages require enclisis to the verb (V=). The same pattern of variation is found between Bulgarian and Macedonian repectively in South Slavic. I refer to this pattern as Mood Swing¹. I argue that in both cases the clitic placement results from adjunction of clitics to C in the (a) but raising of the verb into C in the (b) case. Raising of the verb to C because of mood features seems particularly common in a range of unrelated languages and constitutes a parameter explaining variation. In Minimalism this type of parameter has been interpreted in terms of variation between strong/weak features on heads an approach which requires further theoretical elaboration. An alternative proposed for the South Slavic case is couched in terms of OT, with violable principles of clitic placement in competition with each other (Legendre 1998).
Another issue is whether the parametric difference discussed here, the obligatory raising of V to C in imperatives in some Ngumpin, is related to the obligatory raising of NEG to C in Yapa, and to other features of Ngumpin vs Yapa eg the aspectual nature of Yapa auxiliaries/catalysts versus their modal nature in Ngumpin.
Finally, we explore the relevance of these comparative analyses to the reconstruction of syntactic structures in Proto-NGY and their subsequent change. In the case of the placement of clitics in imperatives and other non-indicative moods, one might posit that NGY had either (a) all clauses 2P; (b) all clauses V=; (c) a split according to mood, retained in some languages, but lost in Yapa. Considerations based on both comparative data from the history of Slavic, Romance etc and theoretical considerations are adduced in searching for an answer.
More generally this kind of alternation is the matrix from which the historical trend which McConvell has called Downward Migration¹ (1996, 2003, 2004) emerges. This is the process of clitics appearing to be move to lower heads in historical sequences, which we argue results from the lower heads (most importantly, verbs) being raised into higher positions such as C. The Mood Swing¹ whereby verbs are raised in non-indicative contexts as in the unrelated subgroups we have discussed provides the background from which categorical clitic placement on the verb (as in some other northern Pama-Nyungan subgroups) can arise.
Languages in contact in the Canberra region in the 19th century. Towards a description of the local Aboriginal language
The recent publication by Ian Clark of George Augustus Robinson's non-Tasmanian journals and Aboriginal vocabularies affords linguists the opportunity to document the Aboriginal language of the Canberra region. Robinson's journal of an 1844 expedition through south-eastern N.S.W. includes a wordlist of some 180 items recorded at Yarralumla, plus considerable information on individual people, including their indigenous names and places of affiliation. To Robinson's documentation can be added a short wordlist from S.M. Mowle, dating from his experiences in the 1830s, plus the evidence of place names taken over by European settlers from the1820s to the 1850s. I will show to what extent it is possible to reconstitute the phonetics and phonology of the language, through the comparison of the same items across different spellings by the same or different recorders of the same or closely related languages, including Curr's wordlists (especially of Queanbeyan and the Monaro) from the 1880s as well as modern professional analyses by Luise Hercus of Southern Ngarrigu and Diana Eades of Dharawal and Dhurga. The relationship of the Canberra language to these and other languages of what Schmidt 1919 called the Yuin language group (including Ngunawal and Gundungurra) will be explored. Conclusions will relate to aspects of methodology, synchronic description, and genealogical relationship.
The relationship between Yalarnnga and Kalkutungu
Kalkutungu and Yalarnnga, two neighbouring languages of northwestern Queensland, are more similar to one another in functional and lexical forms than either is to any other language. In a recent publication Dixon suggests that 'they appear to constitute something resembling a linguistic area' (Dixon 2002: 679). While it is true that there is evidence of convergence, there are also many shared function forms, some obscured by sound changes, which are unlikely to have been borrowed. This raises the question of whether the two languages constitute a relic area or whether they shared an exclusive period of common development, i.e. whether they can be subgrouped.
Comparative Issues in Endophora and Exophora: Jawoyn (southern Arnhem Land) and wider implications
The noun class agreement system in Jawoyn, as in other Kunwinykuan languages (Harvey 1997, Evans 1997), is somewhat loose, with many instances on 'non-standard' agreement. Recognition of this stimulated our closer consideration of demonstratives as part of the larger set of reference-making and cohesive resources in the language.
We consider spatial deixis and textual reference (both to specific items and larger stretches of discourse), and especially the ways in which demonstrative categories overlap in fulfilling these functions, playing a central role in the on-going organization of a communicative frame of reference¹ of which the centrepoint is the current speech situation (see Hanks 2004). We discuss our findings and compare them briefly with Djambarrpuyngu material (Wilkinson 1991), hint at the many parallels with English (Halliday & Hasan 1976), and suggest the further relevance of our approach.
Evans, Nicholas. 1997. Head Classes and Agreement Classes in the Mayali Dialect Chain¹ In M. Harvey & N. Reid (eds.), pp.105-146.
Halliday, M.A.K. & Ruqaiya Hasan. 1976. Cohesion in English. London: Longman Group.
Hanks, William F. 2004. Explorations in the deictic field. ms. Harvey, Mark. 1997. Head and Agreement Classes: An Areal Perspective¹ In M. Harvey & N. Reid (eds)., pp.147-164.
Harvey, Mark & Nicholas Reid (eds.) Nominal Classification in Aboriginal Australia. Amsterdam/Philadelphia: John Benjamins.
Wilkinson, Melanie. 1991. Djambarrpuyngu: a Yolngu variety of Northern Australia. PhD, University of Sydney (ms).
Mon-Khmer Creaky voice: ancient relic or structural tendency? A discussion of the evidence from Bahnaric, Katuic and Vietic
Creaky (or laryngealised) phonation, as a phonemically contrastive feature, is found in a number of Mon-Khmer sub-groups, including Vietic, Pearic, Katuic and North Bahnaric. In some cases, such as the North Bahnaric language Sedang, and the Katuic language Pacoh, it is clear that creaky phonation is a secondary development. However, in other groups, especially some TaOi dialects of Katuic, creaky voice is so well distributed across the lexicon and phonology that it does not appear to be a recent development, and the possibility that it may have to be reconstructed to a much greater time depth has been suggested. However, a clue to the origins of creaky voice in Mon-Khmer can be gleaned from typological considerations: in all cases where creak is attested, so far as I can determine, breathy phonation also exists in languages of the same sub-groups, or existed previously. In the specific case of Sedang we know that a contrast of clear versus breathy phonation shifted to creaky versus clear, in effect the overall system simply became more tense, without any particular segmental conditioning. It is therefore possible to suggest that creaky voice generally arises secondarily out of breathy voice systems, the origins of which are already well understood, being connected with voicing of initial consonants. Segmental features may well affect the distribution of creak, masking the relationship with previously voiced initials, such that all traces of this origin may be lost. More extensive and detailed typological data may well shed considerable light on what is a difficult issue for progress in Mon-Khmer reconstruction.
Recovering linguistic prehistory in the Solomon Islands: the use of complementary methods
Oceania is a region where both traditional and innovative methods of historical linguistics have proven to be useful in recovering linguistic and sociolinguistic prehistory. While the prehistory of some groups of languages in the Pacific can be fittingly explained within the family tree model of language diversification, a model that also includes convergence is needed to explain the prehistory of other groups. This paper takes two subgroups of the Oceanic family, namely Northwest Solomonic and Southeast Solomonic, and looks at how the complementary use of different methods of historical linguistics can lead to a more detailed prehistory of the region than would the use of a single method.
The linguistic boundary between the Northwest Solomonic and the Southeast Solomonic languages is clearly defined by phonological innovations, but there appear to be a number of lexical and grammatical features that are shared across the two groups. In this seminar I will look at a few of these features from the point of view of contact-induced change in an attempt to further add to our knowledge of the linguistic prehistory of the region.
From linguistic to socio-linguistic reconstruction: a study of the 'Kamata' or 'Rajbanshi' lects of north Bengal
The Indo-Aryan languages are commonly said to constitute a dialect-continuum which stretches from Afghanistan to Assam (cf. eg. Masica 1991:25). Such a context presents challenges to historical linguistics that are not unique to Indo-Aryan: How shall we theorise about language genesis so as to do justice to the non-discrete evolution of inter-related lects? Can processes and stages of non-discrete development be reconstructed without marginalising particular processes as non-genetic¹?
This paper responds to these challenges and demonstrates a model I am developing to explain such inter-related genesis among the Indo-Aryan lects of north Bengal. The model is innovative, bringing together historical linguistic, dialectological and sociolinguistic approaches.
Theoretically, the model takes its lead from Milroy, his promulgation of a speaker-based approach to language change (1992, 1999), and the relativisation of the distinction between internally- and externally-induced change the genetic and the areal (Milroy 1997). Following Milroy, Ross (1997) applied the speaker-based approach to the reconstruction of prehistory, developing a theory for using linguistic innovations as windows onto Speech Community Events.
The model outlined in this paper builds on Ross¹ work, to develop a methodology for reconstructing language prehistory within a speaker-based approach. The key methodological innovation is that, rather than focusing exclusively on the reconstruction of linguistic features, the model requires a dual process of reconstruction: concurrently recovering both linguistic features and the linguistic (or dialectological) ranges attained by these features. The model thus brings together spatial and temporal approaches, dialectology and diachrony, without marginalising either.
In the case of north Bengal, where records of social history are available, the recovered changes in linguistic ranges can be compared with historically attested social entities to deliver an inter-disciplinary reconstruction of socio-linguistic history.
Masica, Colin P. 1991. The Indo-Aryan languages. Cambridge: Cambridge University Press.
Milroy, James. 1992. Linguistic variation and change. Oxford: Blackwell.
Milroy, James. 1997. Internal vs external motivations for linguistic change. Multilingua 16(4): 311-323.
Milroy, James. 1999. Towards a speaker-based account of language change. In Ernst Hakon Jahr (ed.) Language change: advances in historical sociolinguistics. Berlin: Mouton de Gruyter.
Ross, Malcolm D. 1997. Social network and kinds of speech-community event. In Roger Blanch and Matthew Spriggs (eds.). Archaeology and Language I: Theoretical and Methodological Orientations. London / New York: Routledge. 1. pp 209-261.
25th June, 2004:
Catastrophic change in current English: Emergent Double-bes and Free-bes
In 1988 I reported in AJL on what was then a relatively new development in English speech what I called double be - at least the first examples I found were from the early 1980's. In the 80's-90's it seems to have spread quickly throughout the (first-language) English-speaking world by diffusion between adults and became well established including in Australia. The kind of sentences in which this classic double-be occurs are equational copular sentences is one in which the subject NP represents the situating in discourse of the predicate, which is a subordinate clause
The main issue is is that they are dividing the party.
The information packaging in this type of clause is unusual in that the subordinate clause is the main assertion and the main clause supplies a discourse frame.
The 1988 paper shows this construction has a particular prosodic pattern which is a blend of two targets which led to the doubling of the copula. It was clearly a new construction and not a performance error. The present paper presents sound data from the US Switchboard corpus which confirms this distinctive character and shows clear diagnostic criteria which separate double be from two types of hesitation-repetition.
In 2004 I heard a radio announcer in Australia say the following:
The headline is is kinda cute
apparently with a classic double be not hesitation-repetition prosody, and I discuss some possible examples in the Switchboard corpus which may be of this extended double be type in normal descriptive sentences with non-clausal predicates. If this is so, this would be a major intrusion of the doubling phenomenon into core grammar with possible radical consequences for the future development of English grammar.
In the 1988 paper I also discussed another emergent grammatical phenomenon involving the copula, which I call free-be which is illustrated by a sentence heard recently on Australian radio
Its important to remember when we discuss the bombing of Japan is that there werent any easy options
I related the emergence of free-be directly to double-be. Recent work on free-be (Ross-Hagebaum 2004) is treating it as a discrete phenomenon but I will argue that in its origin and spread it is indeed strongly linked to the double be.
An activity of the