1) Dialect maps of Harvard survey
During my time as professor of Linguistics at Harvard, I conducted a survey into the various dialects of English spoken in the US, which was concluded in 2003. The survey amounted to 122 questions, mainly investigating phonological differences (e.g. the vowel used in the word "aunt") and lexical variation (e.g. the word for the wheeled contraption used to carry groceries at the supermarket). Each question was multiple choice, meaning that the data could be analysed quantatively; there was also an option for respondents to add extra information in a "comments" box. The aim was to gather data concerning the differences in dialects not only between the North and South states, which has long been a topic of research, but also in certain metropolitan areas and states of the US where research was previously lacking.
Scott Golder, who helped me with the original Harvard survey, plotted the results on maps (both composite and individual), which can be viewed on his website
ABOVE: One of Scott Golder's composite maps for the pronunciation of "caramel" (Harvard Dialect survey Q4)
The survey data and maps are also available at http://www.tekstlab.uio.no/cambridge_survey
, thanks to the help of Marius L. Jøhndal and the University of Oslo. The data has been plotted in exactly the same way as Scott Golder's original maps, but using Google Maps software, which gives the ability to zoom in and out, as well as having labelled states and cities.
ABOVE: An example of the tekstlab Google maps (Harvard Dialect Survey Q50).
2) Analyses / Projects based on Harvard survey
- In 2011, Jack Grieve wrote a paper, entitled "A statistical method for the identification and aggregation of regional linguistic variation". (His complete map set is available here ).
- Joshua Katz (Department of Statistics, NC State University) has also produced several "heat maps" which have received a lot of media attention. See his dialect poster and website for these.
ABOVE: An example of one of Joshua Katz's heat maps (Harvard Dialect Survey Q66) - the intensity of the color corresponds to the dominance of that particular option in the given region.
- "Predicting Hometown Zipcode from Dialect Data" (Emily Tucker, Oregon Graduate Institute)
- "Preliminary Analysis of Dialect Data from the US" (Lifeng Zhu, Centre of Chemometrics, University of Bristol)
- "Application of Chenometrics Approaches to Dialect Data and the Tracing of Mobility of Humans in North America" (Richard G. Brereton, Centre of Chemometrics, University of Bristol)
- "Words in the mind of the individual and the community: De facto standardization of the lexicon" (Neil Wick, Univeristy of Ottawa)
In 2011 or thereabouts, part of my Harvard survey went viral on various social media outlets (youtube and tumblr being the most popular), usually under the name "regional dialect meme", "accent tag" or "accent challenge". As a result, there are countless videos of individuals responding to the survey on the net, creating a mine of information on current English accents - not only from the US, but across the world. I have assembled hundreds of these onto the dialect meme page
of this website so the videos can be accessed more easily.
______________________________________________________________________________________________________3) Subsequent Dialect Surveys
Since conducting the Harvard survey, I have investigated dialects of English outside of the US, e.g. the Survey of English in the British Isles. As a result, comparisons can now be made between the US and the UK, as well as other parts of the world. Data and maps from these subsequent surveys can be viewed on the tekstlab website
ABOVE: An example map (Cambridge Survey Q7) showing the difference between the US and the UK, as well as regional differences in each country