26.3841, Review: Socioling; Text/Corpus Ling: Szmrecsanyi (2015)

Tue Sep 1 00:39:40 UTC 2015

LINGUIST List: Vol-26-3841. Mon Aug 31 2015. ISSN: 1069 - 4875.

Subject: 26.3841, Review: Socioling; Text/Corpus Ling: Szmrecsanyi (2015)

Moderators: linguist at linguistlist.org (Damir Cavar, Malgorzata E. Cavar)
Reviews: reviews at linguistlist.org (Anthony Aristar, Helen Aristar-Dry, Sara Couture)
Homepage: http://linguistlist.org

*****************    LINGUIST List Support    *****************
Please support the LL editors and operation with a donation at:
              http://funddrive.linguistlist.org/donate/

Editor for this issue: Sara  Couture <sara at linguistlist.org>
================================================================

Date: Mon, 31 Aug 2015 20:39:21
From: Luke Rudge [luke.rudge at uwe.ac.uk]
Subject: Grammatical Variation in British English Dialects

Discuss this message:
http://linguistlist.org/pubs/reviews/get-review.cfm?subid=36067437

Book announced at http://linguistlist.org/issues/26/26-1765.html

AUTHOR: Benedikt  Szmrecsanyi
TITLE: Grammatical Variation in British English Dialects
SUBTITLE: A Study in Corpus-Based Dialectometry
PUBLISHER: Cambridge University Press
YEAR: 2015

REVIEWER: Luke A Rudge, University of the West of England

Reviews Editor: Helen Aristar-Dry

SUMMARY 

The linguistic study of British English dialects is one that continues to fascinate and challenge a variety of researchers. From the many instruments available in the linguists’ toolkit, that of ‘corpus-based dialectometry’ (henceforth CBDM) presents perhaps some of the most thought-provoking findings. Extending from the principle that “dialectal data are too complex to be studied one phenomenon at a time” (Nerbonne and Kretzschmar, 2013, p.2), Szmrecsanyi’s approach is indeed one that takes into account a large number of dialectal phenomena from around Great Britain using a sizeable amount of corpus data. In this work, he investigates morphosyntactic features such as modality, negation, agreement and relativisation. In his own words, Szmrecsanyi’s stance is one of looking at “the forest, not the trees” (p.3). Via a series of deductive and inductive methods he is able to present a host of qualitative, quantitative and cartographic results seeking to both reveal more
  information regarding traditional British English dialects and to demonstrate the efficacy of CBDM.

Grammatical Variation in British English Dialects is presented over nine chapters. Chapter 1 introduces the notion of CBDM in an accessible manner, including a brief literature review based on previous endeavours to analyse traditional British English dialects from a broad perspective. Chapter 2 describes the data that is used throughout the book, namely the Freiburg Corpus of English Dialects (FRED), and goes into further detail about the inner-workings in CBDM. Then, thirty-four regions of Great Britain are assessed, each accompanied by an overview of the data held in FRED related to that region, such as word counts and demographic data associated with recordings. Chapter 3 details statistical points of interest within FRED, wherein fifty-seven morphosyntactic features are identified and analysed through an extensive literature review. Through initial analysis, fourteen features with significant geographical distributions are identified. Chapter 4 begins to apply the data and findi
 ngs from Chapters 2 and 3 to numerous maps, providing a visual representation of the ‘distance’ between the dialects of the thirty-four regions, including how closely related each dialect is to both Standard British English and Standard American English. Chapters 5 and 6 inspect these dialectal similarities and differences further in attempting to understand whether dialect distribution is more accurately represented by continua or by clusters, concluding that a mix of the two is more representative of the linguistic landscape. Chapter 7 relocates the study back onto morphosyntactic features, employing various statistical models to find and explain principal features of these dialects, and attempting to explain why certain regions appear anomalous. Finally, Chapters 8 and 9 summarise the study, reasserting the key findings and briefly postulating possible future applications of CBDM whilst comparing it to traditional atlas-based approaches.

EVALUATION 

Szmrecsanyi presents this work in a fashion such that it carries a double function: to investigate morphosyntactic variation in traditional British English dialects, and to provide the reader with a detailed and replicable account of the theories and applications of CBDM. Drawing together a corpus-based approach containing natural speech data with dialectological methodologies, cartography and statistics, this method claims to present a more robust research technique than what has been previously used. While the main arguments in this work are coherent and obvious from start to finish, the explanations of the methodology of CBDM can vary in terms of clarity. For the most part, Szmrecsanyi is able to guide the reader through the thought process by calling upon analogy and exemplifying points using real data from FRED. However, as the book progresses into more statistical territory, the technicality of the writing understandably increases, and the thread of the argument can become lost
  in statistical jargon. These occurrences are nonetheless infrequent and the writing never strays far from the original line of enquiry. References for further investigation are also provided in abundance should the reader be interested in more technical details, including the packages and modules that are implemented in this study via R. Once again, this asserts the position that this book provides guidance for those wishing to replicate or follow this methodology. What aids comprehension the most, though, is access to a variety of colour figures available in Appendix C that visually clarify, for instance, the extent to which certain dialects are (dis)similar to others. A range of maps is called upon: network, beam, honeycomb, similarity, skewness and kernel. Each map is captioned with brief overviews that enable ease of understanding at a glance. When accompanied by their associated texts in the chapters, Szmrecsanyi transmits his ideas and findings in the clearest way possible, a
 lthough the inclusion of some figures (e.g. Map C.1) is questionable given the sheer density of visual information provided in a small area.

Szmercsanyi, like many others in this domain, uses FRED as the data source for this study. From the information made available in Chapter 2, almost 2,500,000 words elicited from over 420 speakers were used in this analysis: an arguably substantial data set. Nonetheless, there are issues and inconsistencies within this data. While the aforementioned figures may seem representative of traditional British English dialects, it is shown that some regions have a much larger share of the total data than others. For instance, the corpus contains 208,000 words from 41 speakers in Somerset, whereas only 6,000 words from 3 speakers are available from Leicestershire. In addition, as this study focuses on traditional dialects, the majority of mean ages for speakers per region falls somewhere between 70 and 85 years. Yet, three regions contain no mean age data for the speakers, and in the case of East Lothian, speakers have a mean age of 17 years. Szmrecsanyi mentions at the start of Chapter 2 tha
 t “we removed some localities with comparatively thin textual coverage from the full corpus” (p.15), rejecting those regions with less than 5,000 words of text available. Thus, a question arises around the inclusion of corpus data wherein the dialect and the subsequent analysis could be affected by unknown or drastically different speaker ages. If these four regions were to be removed from the study, this may produce strong effects upon the resulting calculations, as these require observing the dataset as a whole rather than its parts. The omission of these four regions would result in only a 3% loss in the total word count and a 6% loss in the number of speakers. A remark in Chapter 8 states that results can fit a proposed theory through selection of specific features and rejection of outliers, perhaps similar to the above processes. Yet, to continue with Szmrecsanyi’s arboreal analogy, in cutting down those trees that do not fit the expected trend, we do not provide a repres
 entative picture of the forest. Ergo, any such omissions or modifications must be undertaken carefully and with reason.

Several chapters present findings that concur with the wealth of research performed with traditional British English dialects, such as the North-South dialect divide (see Britain, 2010), yet some findings may surprise the reader. In an extensive review of current literature surrounding the fifty-seven morphosyntactic features identified for use in this study, forty-three are deemed to present insignificant geographical distributions across Great Britain. Rather than following the literature that promotes these features, two statistical tests are applied to the data, leading to a 75% reduction in features deemed significant. In addition, Table 8.2 (p.158) summarises findings of this study with similar studies, identifying how well each approach explains geographical differences in dialect. High values are attributed to atlas-based studies focussing on pronunciation, yet the values for Szmrecsanyi’s work are comparatively small. Hypotheses are provided to explain the differences in t
 hese values, resulting in the conclusion that atlas-based methodologies are inaccurate. Séguy’s Law could be used to back up this claim: “after a certain threshold, geography is not supposed to matter that much any more” (p.102). What is not called into question, however, is the methodology employed in this study. Rather than only identifying the flaws in other studies, a reflexive approach may need to be performed in order to re-evaluate and scrutinise the methods within CBDM. This is performed in certain sections, such as the discussion that ensues concerning choice between two algorithms available for inductive analysis (p.118). Thus, it may be worthwhile performing further reflections throughout the full CBDM methodology to ensure that it remains a robust approach.

These comments should not be viewed as dismissals of this thorough investigation or of CBDM as a systematic methodology. Szmrecsanyi briefly mentions further possibilities that this approach possesses, ranging from deeper investigation into morphosyntactic variation, to examinations of lexical and phonological variation between dialects. Studies such as these would allow for a more detailed scrutiny of older atlas-based methods, especially with regards to reaffirming or refuting claims that these methods “overestimate the importance of geography” (p.160) in their findings. It would also be worthwhile to apply CBDM to further corpora, such as the Helsinki Corpus of British English dialects and the Salamanca Corpus, to observe how different data may validate or refute the claims made in this study. Finally, CBDM could be employed to observe other languages of Great Britain, such as British Sign Language. Recent research has supported the claim of regional levelling in this visual-s
 patial language (see Stamp, Schembri, Fenlon, Rentelis, Woll and Cormier, 2014), and as the British Sign Language Corpus develops, CBDM would prove a useful tool in examining the variation in this language. If these future studies are performed and recorded in the style of Szmrecsanyi, it will undoubtedly provide an accessible and thought-provoking piece of work for students of linguistics and professionals alike.

REFERENCES

Britain, David. 2010. Grammatical variation in the contemporary spoken English of England. In Kirkpatrick, Andy (ed.). The Routeledge Handbook of World Englishes, 37-58. London: Routeledge.

Nerbonne, John & William A. Kretzschmar Jr. 2013. Dialectometry++. Literary and Linguistic Computing, 28(1), 2-12. doi:10.1093/llc/fqs062

Stamp, Rose, Adam Schembri, Jordan Fenlon, Ramas Rentelis, Bencie Woll & Kearsy Cormier. 2014. Lexical variation and change in British sign language. PLoS ONE 9(4), 811–824. doi:10.1371/journal.pone.0094053

ABOUT THE REVIEWER

Luke Rudge is a Ph.D research student at the Bristol Centre for Linguistics at the University of the West of England. Although having broad interests across linguistic disciplines, his research currently specialises in sign linguistics, language variation and functional language analysis. In approaching the end of his first year of doctoral study, he soon hopes to complete his thesis and move into lecturing. Luke also sits on the United Kingdom Linguistics Olympiad committee, assisting schools and colleges across the UK with setting up and running Linguistics Clubs for pupils and students.

----------------------------------------------------------
LINGUIST List: Vol-26-3841	
----------------------------------------------------------