Linguistictypology is an approach to the sci- entific study of language which was pioneered in its modern form by Joseph Greenberg in the 1950s and 1960s (see e.g. Greenberg, 1963). 8 In the in- tervening decades, it has evolved from a search for language universals and the limits of language variation to what Bickel (2007) characterizes as the study of “what’s where why”. That is, typol- ogists are interested in how variations on particu- lar linguistic phenomena are distributed through- out the world’s languages, both in terms of lan- guage families and geography, and how those dis- tributions came to be the way they are.
What we will present in this paper is a generative model that corresponds to a generative tradition of research in linguistictypology. We first outline the technical linguistic background necessary for the model’s exposition. Chomsky famously argued that the human brain contains a prior, as it were, over possible linguistic structures, which he termed universal grammar (Chomsky, 1965). The con- nection between Chomsky’s Universal Grammar and the Bayesian prior is an intuitive one, but the earliest citation we know for the connection is Eisner (2002, §2). As a theory, universal grammar holds great promise in explaining the typological variation of human language. Cross-linguistic similarities and differences may be explained by the influence universal grammar exerts over language acquisition and change. While universal grammar arose early on in the writtings of Chomsky, early work in generative grammar focused primarily on English (Harris, 1995). Indeed, Chomsky’s Syntactic Structures contains exclusively examples in English (Chomsky, 1957). As the generative grammarians turned their focus to a wider selection of languages, the principles and parameters framework for syntactic analysis
The volume starts with a typological introduction outlining the marking, and the meaning, of evidentials and other ways of marking information source, together with cultural and social aspects of the conceptualization of knowledge in a range of speech communities. It is followed by revised versions of twelve of the fourteen presentations from the International Workshop 'The grammar of knowledge', held at the Language and Culture Research Centre, James Cook University, 16-21 July 2012. An earlier ver sion of Chapter 1 had been circulated to the contributors, with a list of points to be addressed, so as to ensure that their detailed studies of individual languages were cast in terms of a common set of typological parameters. (This is the seventh monograph in the series Explorations in LinguisticTypology, devoted to volumes from the Inter national Workshops organized by the co-editors.)
Simon Overall received his MA in Linguistics from the University of Auckland in 1999. and in 2007 he completed his Ph.D. at RCLT, La Trobe University. His Ph.D. dissertation was a grammar of Aguaruna, a Jivaroan language of the northern Peruvian Amazon. Currently he is a Postdoctoral Research Fellow at RCLT, continuing his research into the Jivaroan languages and investigating their genetic and areal relations with neighbouring Amazonian and Andean languages. Address. Research Centre for LinguisticTypology, La Trobe University, Victoria 3086, Australia; e-mail: firstname.lastname@example.org.
The field of linguistictypology studies the similarities and distinguishing features between languages and aims to classify them accordingly. Among other areas, the World Atlas of Language Structures describes general properties of each language’s word order. Overall, WALS contains 192 features, but not all features are relevant to determining word order. Many WALS features deal with phonology, morphology or lexical choice: Feature 129A, for example, describes whether the language’s words for “hand” and “arm” are the same. Hence, for simplicity’s sake we pre-select the subset of WALS features potentially relevant to determining word order and describe this subset in the following. Table 1 provides an overview of these features, along with an indication of the relative frequency distribution of each of their values over all languages in WALS.
Typological features have two advantages over other linguistic traits. First, they allow us to com- pare an arbitrary pair of languages. By contrast, historical linguistics has worked on regular sound changes (see (Bouchard-Cˆot´e et al., 2013) for com- putational models). Glottochronology and computa- tional phylogenetics make use of the presence and absence of lexical items (Swadesh, 1952; Gray and Atkinson, 2003). All these approaches require that certain sets of cognates, or words with common et- ymological origins, are shared by the languages in question. For this reason, it is hardly possible to use lexical evidence to search for external relations in- volving language isolates and tiny language families such as Ainu, Basque, and Japanese. For these lan- guages, typology can be seen as the last hope.
The Stanford Dependencies (SD) representation (de Mar- neffe et al., 2006) was originally developed as a practical representation of English syntax, aimed at natural language understanding (NLU) applications. However, it was deeply rooted in grammatical relation-based syntactic traditions, which have long emphasized cross-linguistic description. Faithfulness to these origins was attenuated by desiderata from our NLU applications and the desire for a simple, uni- form representation, which was easily intelligible by non- experts (de Marneffe and Manning, 2008). Nevertheless, it is reasonable to suppose that these additional goals do not detract from cross-linguistic applicability.
As reviewed above, HG fundamentally differs from OT because it does away with strict domination and therefore allows for gang effects in which multiple violations of lower-weighted constraints outweigh a violation of a higher-weighted constraint (see sec- tion 3 for details). Bane and Riggle (2009) show that sets of constraints drawn from the phonological literature yield much richer typologies in HG than in OT as a result of gang effects, and that many of the additional patterns derived under HG are unat- tested. The same point is made by the investigation of Kaun’s (2004) analysis of the typology of round- ing harmony discussed in section 4. However these constraint sets were developed in the context of OT, so these results leave open the possibility that a re- vised HG constraint set could provide a closer match
In the experiments for RQ1 and RQ2 we predict typological features extracted from WALS (Dryer and Haspelmath, 2013). We choose to investi- gate three linguistic levels of language: phonol- ogy, morphology, and syntax. This is motivated by three factors: (i) these features are related to NLP tasks for which data is available for a large lan- guage sample; (ii) the levels cover a range from basic phonological and morphological structure, to syntactic structure, allowing us to approach our research question from several angles; and (iii) the features in these categories are coded in WALS for a relatively large selection of languages. We ex- tract the three feature sets which represent these typological levels of language from WALS. 1
The Amazon basin is the least known and least understood linguistic region in the world. Maps of the language families of South America (with one colour for each genetic group) purvey an impression of anarchy – there are dabs of yellow and blue and red and orange and brown mingled together like a painting by Jackson Pollock. And when one does get hold of a grammar of an Amazonian language it is likely to show strange properties – multiple sets of classifiers, oddly conditioned ergativ- ity splits, and so on – that constitute exceptions to received ideas about typological universals. In other instances one finds the richest examples of categories that are weakly attested elsewhere. For instance, Tucano languages (chapter 7) have the most highly articulated systems of evidentiality in the world; this is an obligatory specification of the evidence a speaker has for making a statement – whether observed, or reported, or inferred, or assumed. However, a major di ffi culty is that a high proportion of available grammars are incomplete, a ff ording a glimpse of some exotic grammatical property but with insu ffi cient information to enable the reader to fully understand it, and to realize its overall typological significance.
Even t hough f rag mented desig n of European airspace is quite a recognizable problem (which is frequently mentioned w ith in t he ATM com munit y), sti l l there are many unanswered questions requiring comprehensive analysis with a view to provide appropriate answers and improvement guidelines. By reviewing literature this research aims to define European airspace fragmentation typology and to answer following research questions - Which fragmentation types exist?; What are the differences between them?; How they can be displayed?; Which types are more apparent than others?; How much do they spatially and temporally vary?; When did some types began to appear?; Which fragmentation types have been studied the most?; and finally Which fragmentation types require further research?