Meet Our Keynote Speakers


Tony Berber Sardinha Catholic University of Sao Paulo, Brazil

Tony Berber Sardinha is a professor with the Graduate Program in Applied Linguistics and the Linguistics Department, the Catholic University of Sao Paulo, Brazil. His recent publications include Multidimensional Analysis: 25 Years on (2014, John Benjamins), Working with Portuguese Corpora (2015, Bloomsbury), Metaphor in Specialist Discourse (2015, John Benjamins), and Multidimensional analysis (forthcoming, Routledge). His main research interests are multi-dimensional analysis, the use of corpora for historiography and historical discourse analysis, the development of corpus methods for metaphor retrieval, the application of corpus techniques in forensic linguistics, corpus-based analysis of translated texts, and the interface between corpus linguistics and language teaching. He is on the board of several journals and book series such as the International Journal of Corpus Linguistics, Corpora, Metaphor and the Social World, Studies in Corpus Linguistics, Register Studies, and Metaphor in Language, Cognition, and Communication.



 Topic: Corpus Linguistics and Historiography

Historiography can mean both the development of principled approaches for the study of history and the study of history based on such approaches. The goal of this talk is two-fold: first, to make the case that corpus linguistics is well-suited to engage with historiography and, second, to present a study that demonstrates corpus linguistic historiography. The historical accounts derived from a corpus perspective are entirely text based—that is, they represent the discourses that actually emerge from the texts, mined directly from primary sources. The principal research questions asked in this kind of investigation include: What are the different discourses that exist in the field? How are they constituted? How do these discourses vary over time and across the publications in the field? To answer such questions, corpora comprising as long a time frame as possible need to be collected. The corpus used in the study covers a 70-year span of publications, containing the full collection of papers published in major journals. The corpus was tagged for part of speech and analyzed by means of a lexical multi-dimensional analysis. Three multivariate statistical procedures were employed in the analysis of the corpus. First, a factor analysis was used to determine the groupings of lexical items cooccurring across the texts. These factors were interpreted in terms of such constructs as lexical fields, topic selection, and semantic preference, thereby giving rise to the lexical dimensions underlying the variation among the texts over time. The second procedure was a discriminant function analysis, which was employed to test the robustness of the dimensions as predictors of the time periods in which the texts were published. The third procedure was cluster analysis, which was run on the dimension scores in order to group the texts around the prevailing discourses in the field. Finally, a timeline was derived from the cluster analysis, reflecting the major time periods in the discourse of applied linguistics over the course of its history.


Bethany Gray Iowa State University, Ames, Iowa

Bethany Gray is Associate Professor of English at Iowa State University (program in Applied Linguistics and Technology). Her research uses corpus linguistics methods to explore register variation in English, with a particular focus on the phraseological and grammatical characteristics of texts, English for Academic Purposes and disciplinary variation, and second language writing development. Her work has appeared in journals such as Applied Linguistics, International Journal of Corpus Linguistics, Corpora, TESOL Quarterly, Journal of English for Academic Purposes, English for Specific Purposes, Journal of English Linguistics, and English Language & Linguistics, and has been published in book form through Cambridge University Press and John Benjamins. She is a co-founding editor of the          journal Register Studies (John Benjamins).


 Topic: The Role of Register Analysis in Academic Writing Research

Register analysis provides a framework for the systematic consideration of situational factors that underlie functional linguistic variation across texts. This presentation explores the importance of register analysis as a key component of corpus-based research on the linguistic characteristics of academic writing, arguing that such research can benefit from greater register awareness and specificity. Analyses of grammatical variation in expert and L2 production demonstrate how an explicit consideration of registers and sub-registers within academic writing has implications for corpus design and representativeness, corpus-based research design, investigations of disciplinary variation, and studies of writing development.




Stefan Th. Gries University of California, Santa Barbara and Justus Liebig University

Giessen Stefan Th. Gries (Full) Professor of Linguistics in the Department of Linguistics at the University of California, Santa Barbara (UCSB) and a 25% part-time Chair of English Linguistics (Corpus Linguistics with a focus on quantitative methods) and Honorary Liebig-Professor of the Justus-Liebig-Universität Giessen. He earned his M.A. and Ph.D. degrees at the University of Hamburg, Germany, in 1998 and 2000 and started teaching/research at the University of Southern Denmark at Sønderborg (1998–2005). After a brief stay in the Psychology Department of the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, he accepted a position at UCSB, starting November 1, 2005. Gries was a visiting professor at the 2007, 2011, 2013, and 2015 LSA Linguistic Institutes at Stanford University, the University of Colorado at Boulder, the University of Michigan, Ann Arbor, and the University of Chicago as well as a Visiting Chair (2013–2017) of the Centre for Corpus Approaches to Social Science at Lancaster University and the Leibniz Professor (spring semester 2017) at the Research Academy Leipzig of Leipzig University.


Topic: Towards a Unified Tupleization of Corpus Linguistics

In a sense, this is the final talk in a series of three plenary talks this year on how corpus linguistics should change its quantitative analyses of corpus data. In two previous talks, I argued that corpus linguistics is held back by our (often implicit) decision(s) to not utilize all the different dimensions that corpus data have to offer and to use measures that sometimes even mask information that would be relevant to our analyses. In those previous talks, I then proceeded by discussing dimensions of information that we need to use more (e.g., variability and/or dispersion) and measures that are better suited to providing us with the relevant information (e.g., directional association measures and/or entropy).In this talk, I will try and, to say it with a great amount of hubris, exemplify how I think we should be doing corpus statistics. More specifically, the field is teeming with many different measures for different purposes (there dozens of association measures, more than a dozen dispersion measures, many association measures are also used for key word statistics, …) that suffer from some of the problems alluded to above, that are diverse and often not particularly comparable, and that defy a unified perspective on the kinds of distributional data we are always working with. I will therefore propose and of course exemplify an alternative statistic that- can be applied to most standard uses in/with corpus-linguistics statistics;- avoids many shortcomings existing measures exhibit;- has a solid theoretical grounding utilized across a huge number of other disciplines (including cognitive science);- allows us to keep different pieces of information that existing approaches routinely and unhelpfully conflate separate;- even offers new avenues of research.




Click here to return to AACL 2018 home page