Language in Space

Geographic Perspectives on Language Diversity and Diachrony

Andrea L. Berez
University of California Santa Barbara/University of Hawaii at Manoa

Lexical and Discourse Vividness in Ahtna Directionals: Observations Using GIS

Travel narration is a genre of Ahtna (Athabascan) spontaneous monologue that is characterized by spatially rich descriptions of overland foot travel covering hundreds of square miles of Ahtna territory in southcentral Alaska. Among the grammatical systems used in the narratives to describe travel paths is a class of semantically riverine directionals. The directionals have a historically complex morphological structure that is undergoing lexicalization, causing individual morphemes to become semantically bleached and lose their capacity to describe precise and subtle differences in direction and location. This paper uses Geographic Information Systems technology and high-resolution topographic imagery to examine the effects of lexicalization and speakers' subsequent discourse strategies for vividly describing directional concepts that are necessary for travel narration.


Mark Donohue and Bronwen Whiting
Australian National University

Kinds of Lexical Similarity

While 'simple' lexicostatistics is often uninformative, or open to too many alternative analyses, a more detailed examination of wordlists by semantic fields is intuitively more attractive to uncover the effects of geography and contact on the lexicon of a language. Examining a published set of wordlists from central Indonesia, we divided them into different semantic fields, such as body parts, kinterms, etc., and coded the different languages by absolute and relative distance. This allowed us to examine which of the different semantic fields covary to a significant extent, and so can be treated as single data types, and which were independent. We could also determine the effects of relatedness and (social) distance on the lexical similarities between varieties. While a general distance decay effect is observed in the signal, and predicted by many earlier studies, we also found negative correlations with distance for some semantic fields.


Russell Gray
University of Auckland

On the Pleasures and Perils of Phylolinguistics

The use of computational phylogenetic methods in historical linguistics is growing quite rapidly. However, the utility of these methods and the accuracy of recent high profile results are the subjects of considerable dispute. In this talk I will outline how Bayesian phylogenetic methods, including very recent developments in phylogeographic modeling, can shed light on long standing questions such as the origin of the Indo-European languages and the timing of the Austronesian expansion. I will emphasize that these methods should be seen as a supplement to, rather than a replacement of, traditional linguistic scholarship.


Hannah J. Haynie
University of California Berkeley

The Geographic Dimension of Linguistic Diversity within the Eastern Miwok Family

The Eastern Miwok family is commonly characterized as a set of four related, geographically-named languages (Northern Sierra Miwok, Central Sierra Miwok, Southern Sierra Miwok, and Plains Miwok) spoken across the western Sierra Nevada slopes and into the Sacramento River Delta. This talk re-examines the internal diversity of Eastern Miwok, using isogloss mapping to evaluate this family's standard internal classification and language boundaries. An ecological approach, examining spatial patterns in linguistic diversity in the context of physical landscape features and material culture regions, allows us not only to investigate where particular patterns of dialect diveristy occur but also why they arose.


Gary Holton
University of Alaska Fairbanks

A Geo-linguistic Approach to Understanding Relationships within the Athabaskan Family

It is widely acknowledged that the Athabaskan languages do not lend themselves to subgrouping according to a traditional Stammbaum or family-tree model (Sapir 1915, Hoijer 1963, Krauss 1973). Rather, Athabaskan is characterized by cross-cutting phonological, morphological and lexical developments which make it difficult to identify subgroups based on shared innovations. Nevertheless, numerous ad-hoc subgroupings have been proposed, none of them offering much explanatory value. This paper proposes an alternative approach to understanding Athabaskan genealogical relationships, based on mapping of individual isoglosses. These isoglosses are for the most part well-established, but modern mapping techniques allow them to be viewed in more productive ways. By layering isoglosses we can easily visualize historical developments which have shaped the Athabaskan region.


Brian Kemp
Washington State University

Evaluating the Farming/Language Dispersal Hypothesis with Genetic Variation Exhibited by Populations in the Southwest and Mesoamerica

The Farming/Language Dispersal Hypothesis posits that prehistoric population expansions, precipitated by the innovation or early adoption of agriculture, played an important role in the uneven distribution of language families recorded across the world. In this case, the most widely spread language families today came to be distributed at the expense of those that have more restricted distributions. In the Americas, Uto-Aztecan is one such language family that may have been spread across Mesoamerica and the American Southwest by ancient farmers. We evaluated this hypothesis with a large-scale study of mitochondrial and Y-chromosomal DNA variation in indigenous populations from these regions. Partial correlation coefficients, determined with Mantel tests, demonstrate that Y-chromosome variation in indigenous populations from the American Southwest and Mesoamerica correlates significantly with linguistic distances (r=0.33-0.384, p<0.02), while mitochondrial DNA (mtDNA) diversity correlates significantly with only geographic distance (r=0.619, p=0.002). The lack of correlation between mtDNA and Y chromosome diversity is consistent with differing population histories of males and females in these regions. However unlikely, if groups of Uto-Aztecan speakers were responsible for the northward spread of agriculture and their languages from Mesoamerica to the Southwest, this migration was possibly male biased. However, a recent in situ population expansion within the American Southwest (2105 years before present (YBP) [99.5% CI 1273-3773 YBP]), one that probably followed the introduction and intensification of maize agriculture in the region, may have blurred ancient mtDNA patterns, which might otherwise have revealed a closer genetic relationship between females in the Southwest and Mesoamerica.


Patrick McConvell
Australian National University

Language Spread by Australian Hunter-Gatherers in Prehistory: Detecting and Modeling Language Shift

Language spread by farmers and pastoralists has attracted most attention, especially as part of an effort to explain widespread language families as resulting from 'demic diffusion' generated by rapid population growth underpinned by agriculture., This paper focuses on the fact that hunter-gatherer languages have also spread widely, to form large language families. The list of such widespread hunter-gatherer language families still observable today is quite long (e.g Pama-Nyungan in Australia).
Among hunter-gatherers we are interested in the mechanisms which spread the languages. All language spread has a migration component, either migration pure and simple into empty or sparsely populated zones, or by language shifts to the language of migrants, known as language replacement. ('upstream' vs. 'downstream spread', respectively: McConvell 2001, 2010).
In this paper the main topic will be language shift, which is more characteristic of downstream spread. Two issues will be examined: (1) how can language shift be detected in prehistory -- what kind of evidence from linguistics in particular can establish that a shift has taken place? (2) how can we model language shift in hunter-gatherer populations to achieve a better understanding of its processes?
The focus here is on specific kinds of linguistic evidence which may point to prior language shift: 1. The hypothesis of Thomas on ad Kaufman (1988) that phonology and syntax are the prime areas which are subject to substratal interference. 2. In lexicon, that environmental vocabulary is derived from substrate languages particularly where the new language is moving into a new environmental zone. 3. Some place names are adopted from the substrate when language shift takes place.
In all these cases it is important to provide evidence for why the effects are probably due to shift rather than adstratal contact and borrowing.
I examine a case of probable hunter-gatherer language shift in prehistory, the spread of the Eastern Ngumpin languages into the Victoria River District in the Northern Territory of Australia. I also look at cases where the mechanism of language spread was probably mainly migration without significant language shift ('upstream'), such as the Western Desert language. In some areas of the Western Desert, however, there is a significant amount of effects 2 and 3. In this case however the most recent eastern spread of the language may have been 'downstream', having a language shift component.


John Nerbonne
University of Groningen

Aggregate vs. Feature-based Perspectives on Dialect Geography

Dialect geography studies the distributions of linguistic variation of different sorts and different levels of granularity. Traditional studies emphasized the study of individual features, and dialectometry has emphasized aggregate differences (Goebl 1984, Nerbonne 2009), offer arriving at similar conclusions. It is interesting to explore the nature of the link between the two in order to appreciate how and whether they complement each other (Wieling and Nerbonne, 2011). I'll report on several techniques that aim to extract most individual features and how they seem to be yielding similar results. I will be reporting on joint work with Martijn Wieling, Harald Baayen, Bob Shackleton and Jelena Prokic.


Johanna Nichols
University of California Berkeley

Open and Closed Spread Zones

A spread zone is an area where, ultimately for geographical reasons, languages tend to spread out widely, driving previous languages in the area to extinction, so that when a spread is complete only one language occupies most or all of the area. Spread zones occur in continental interiors and dry or seasonal climates (including arctic and subarctic). Though diversity in a spread one is low, successive spreads bring diversity over time and cause a diverse set of survivors of previous spreads to build up at the periphery of the spread zone.
This view of diversity proves accurate only for open spread zones, ones that can be entered from several or all sides. Examples include the North American Great Plains, the U.S. Great Basin, the California central valley, and the distributed oasis spread zone of Central Asia.
A closed spread zone is surrounded by geographical or ecological barriers, so that once a language has spread through the area it is unlikely to be supplanted by a new entrant; therefore successive spreads involve daughters or neighbors of the previous spreading language, and diversity decreases overall, even in the periphery. Examples are Australia, Africa, the eastern and Mongolian steppes in Eurasia, and mountain altiplano areas (Tibet, the Andes, the New Guinea highlands). In each an eccentric typological profile results from isolation, and where the spread zone is either large or old it can have a disproportionate effect on continental or worldwide typological frequencies. In all of them the major factor in their typological and genealogical formation has been extinction. The paper reviews the histories of these closed systems and their larger impact.


Loretta O'Connor
Radboud University

Languages of the Intermediate Area in Conceptual and Physical Space

Part of the geographic region that connects the American continents was dubbed by archaeologists as "the Intermediate Area" to reflect its position between the better-studied cultural zones of Mesoamerica and the Andes. This territory stretches from northern Honduras through the isthmus of Panama and into the northwestern areas of Colombia, Venezuela and Ecuador.
The many languages and language families of the Intermediate Area have proven a puzzle for linguists engaged in defining their genetic and contact histories. A case in point is the meshwork of partially overlapping proposals related to the Chibchan, Chocoan, Barbacoan and Paesan languages. This study brings together structural and lexical data from these families and selected neighbors for analysis with two types of computational tools -- phylogenetic estimation methods and GIS -- to propose alternative explanations of relationship between and among these groups.


Sergio Rey
Arizona State University

Exploratory Analytics for Spatial Dynamics with PySAL and STARS

This talk provides an overview of recent work on exploratory space time data analysis. These methods can be placed into one of three classes. First are those that have taken exploratory spatial data analysis methods designed for cross-sectional settings and extended these to a dynamic context. The second class of approaches begin with what are dynamically oriented analytics that are changed to include a spatial component. These first two classes can be thought of as conditional in the sense that they provide insights as to the role of one dimension (space or time) conditioned on the state of the other dimension. The third group of methods are unconditional and are designed to consider the joint space-time dimensions of the data. The implementation of a set of these methods in the open source packages PySAL (Python Spatial Analysis Library) and STARS (Space-Time Analysis of Regional Systems) is also illustrated.


Benedikt Szmrecsanyi
Freiburg Institute for Advanced Studies

Modeling Dialect Diffusion and Diversity on the Basis of Naturalistic Speech Data

This paper introduces methodologies to tap naturalistic speech corpora for exploring linguistic distances between dialects or varieties as a function of properties of geographic space. My case study draws on the Freiburg Corpus of English Dialects, a major dialect corpus that samples interview material from more than thirty traditional English dialects all over Great Britain. On the basis of text frequencies of several dozen morphosyntax features, the study calculates a measure of aggregate linguistic distance. This measure is subsequently analyzed to uncover patterns of areal diffusion and regionalization, and to probe the explanatory power of language-external predictor variables such as least-cost travel time or linguistic gravity.