Linguistic Anthropology - Social Sciences

1 Statistics based on the full quantile regression have been several proposals dealing with gener-alizations of quantile regression to nonparametricresponse functions involving both local polynomialmethods and splines. Extension of quantile regressionmethods to multivariate response models is a par-ticularly important ConclusionClassical least squares regression may be viewed as anatural way of extending the idea of estimating anunconditional mean parameter to the problem ofestimating conditional mean `functions,' the crucialstep is the formulation of an optimization problemthat encompasses both problems. Likewise, quantileregression offers an extension of univariate quantileestimation to estimation of conditional quantile func-tions via an optimization of a piecewise linear objectivefunction in the residuals.

2 Median regression minimizesthe sum of absolute residuals, an idea introduced byBoscovich in the eighteenth asymptotic theory of quantile regression closelyparallels the theory of the univariate sample quantiles;computation of quantile regression estimators may beformulated as a linear programming problem andefficiently solved by simplex or barrier methods. Aclose link to rank based inference has been forgedfrom the theory of the dual regression quantile process,or regression rankscore nontechnical introductions to quantile re-gression are provided by Buchinsky (1998) andKoenker and Hallock (2001). Most of the majorstatistical computing languages now include somecapabilities for quantile regression estimation andinference. Quantile regression packages are availablefor R and Splus from the R archives at http:}} }R}CRAN and Statlib at http:}} }S, respectively.

3 Stata's central coreprovides quantile regression estimation and inferencefunctions. SAS offers some, rather limited, facilitiesfor quantile I, Roberts F D K 1974 Solution of an overdeterminedsystem of equations in thel" ACM17:319 20 Buchinsky M 1998 Recent advances in quantile regressionmodels: a practical guide for empirical ofHuman Resources33: 88 126 Doksum K 1974 Empirical probability plots and statisticalinference for nonlinear models in the two sample Statistics2: 267 77 Gutenbrunner C, Jurec)kova!J 1992 Regression quantile andregression rank score process in the linear model and of Statistics20: 305 30Ha!jek J, Sida!k Z 1967 Theory of Rank Tests. Academia, Prague,Czech RepublicHyndman R J, Bashtannyk D M, Grunwald G K 1996 Es-timating and visualizing conditional Graphical Statistics5: 315 36 Koenker R, Bassett G 1978 regression : 33 50 Koenker R, Hallock K 2001 Quantile ofEconomic,Perspecti es, forthcomingLehmann E 1974 Nonparametrics: Statistical Methods Based onRanks.

4 Holden-Day, San FranciscoManski C 1985 Semiparametric analysis of discrete response:asymptotic properties of the maximum score of Econometrics27: 313 34 Portnoy S, Koenker R 1997 The Gaussian hare and theLaplacian tortoise: computability of squared-error estimators, with Science12: 279 300 Powell J L 1986 Censored regression ofEconometrics32: 143 55R. KoenkerLinguistic Anthropology ` Linguistic Anthropology ' is an interdisciplinary elddedicated to the study of language as a culturalresource and speaking as a cultural practice. It assumesthat the human language faculty is a cognitive and asocial achievement that provides the intellectual toolsfor thinking and acting in the world. Its study must bedone by detailed documentation of what speakers sayas they engage in daily Social activities.

5 This docu-mentation relies on participant observation and othermethods, including audiovisual recording, annotatedtranscription, and interviews with an interdisciplinary eld, Linguistic anthropologyhas often drawn from and participated in the de-velopment of other theoretical paradigms. Some of itsown history is re ected in the oscillation often foundamong a number of terms that are not alwayssynonyms: Linguistic Anthropology , anthropologicallinguistics, ethnolinguistics, and sociolinguistics. Itsmain areas of interest have changed over the years,from an almost exclusive interest in the documentationof the grammars of aboriginal languages to theanalysis of the uses of talk in everyday interaction andthroughout the life span (Duranti 1997, Foley 1997).

6 This article provides a brief historical account oflinguistic Anthropology , and highlights important pastand present issues, theories, and Linguistic Anthropology within the BoasianTraditionIn the holistic tradition established by Franz Boas(1858 1942) in the USA at the beginning of thetwentieth century, Anthropology was conceived as8899 Linguistic Anthropologycomprising four sub elds: archaeology, physical (now`biological') Anthropology , linguistics (now `linguisticanthropology'), and ethnology (now `socioculturalanthropology'). This vision of Anthropology differsfrom the one found in the European tradition, wherelinguistics and Social Anthropology remained rigidlyseparate disciplines for most of the twentieth century,despite the emphasis on the use of native languages in eldwork among UK anthropologists, and the theor-etical and methodological in uence of BronislawMalinowski (1884 1942), who wrote about the im-portance of Linguistic research for an anthropologicalunderstanding of human societies.

7 In the 1950s, theadoption of the term `ethnolinguistics' (re ecting theEuropean preference for `ethnology' over `(cultural) Anthropology ') for those studies that merged linguisticand anthropological interests signaled the intellectualrecognition, at least in some European academiccircles, of the importance of an `ethnological side' oflinguistic studies (Cardona 1976), but the institutionalrecognition of such a discipline within Europeananthropology has been slow to come. Europeanscholars with research interests similar to those ofNorth American Linguistic anthropologists are thusmore likely to be found in departments of linguistics,foreign languages and literatures, folklore, communi-cation, sociology, or understand the special role given to the study oflanguages in the Boasian tradition, we must go back tothe time when Anthropology became a profession inthe USA, in the period between the last decades of thenineteenth and the rst decades of the twentiethcentury.

8 In that time, the study of American Indianlanguages emerged as an essential part of anthro-pological research. John Wesley Powell (1834 1902),the founder of the Bureau of Ethnology, later renamedBureau of American Ethnology (BAE), supported,through grants from the US government, Linguistic eldwork, in the belief that by collecting vocabulariesand texts from American Indian languages, it wouldbe possible to reconstruct their genetic relations andthus help in the classi cation of American Indiantribes. Boas himself had become fascinated by thegrammatical structures of Chinook and otherlanguages of the American northwest coast early on inhis eldwork, and seized the opportunity to work forthe BAE and edit theHandbook of American IndianLanguages(1911).Although Boas, a diffusionist, was quite skeptical ofthe possibility of using languages for reconstructinggenetic relations between tribes and was against anycorrelation between language and race he tried totransmit to his students a passion for the details oflinguistic description and the conviction thatlanguages were an important tool for (a) eldwork,and (b) the study of culture, especially because thecategories and rules of language are largely uncon-scious and thus not subject to secondary rational-izations.

9 Furthermore, Boas was committed to whatlater became known as `salvage Anthropology ,' that isthe documentation of languages and cultural tra-ditions that seemed on the verge of disappearing. Thisenterprise a struggle against time due to the greatdamage already done, by the end of the nineteenthcentury, to the indigenous cultures of the Americas byEuropean colonizers produced valuable informationon Native American traditions, but it had its method-ological and theoretical drawbacks, the most agrantof which was the inability to see or accept the effects ofcultural contact and his writing and teaching, Boas broughtscienti c rigor to Linguistic description and helpeddemolish a number of unfounded stereotypes aboutthe languages that were then called `primitive.' In an1889 article entitled `On alternating sounds,' Boasargued that the commonly held view that speakers ofAmerican Indian languages were less accurate in theirpronunciation than speakers of Indo-Europeanlanguages was false and probably due to the lack oflinguistic sophistication of those who had rst triedto describe indigenous languages.

10 Consistent with hiscultural relativism, Boas believed that each languageshould be studied on its own terms rather thanaccording to some preset categories based on the studyof other, genetically unrelated languages ( , Latin).In his `Introduction' to theHandbook(1911), Boasprovided an overview of the grammatical categoriesand Linguistic units necessary for the analysis ofAmerican Indian languages and argued against over-generalizations that would obscure differences acrosslanguages. He identi ed the sentence (as opposed tothe word) as the unit for the expression of ideas, andlisted a number of grammatical categories that arelikely to be found in all languages, while pointing outthat the material content of words (the meaning oflexical items) is language-speci c and that languagesclassify reality differently.