BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//University of Liverpool Computer Science Seminar System//v2//EN
BEGIN:VEVENT
DTSTAMP:20260921T211006Z
UID:Seminar-DMML-1109@lxserverM.csc.liv.ac.uk
ORGANIZER:CN=Danushka Bollegala:MAILTO:Danushka.Bollegala@liverpool.ac.uk
DTSTART:20210202T110000
DTEND:20210202T120000
SUMMARY:Data Mining and Machine Learning Series
DESCRIPTION:Emma Moreton: New methods, old data: exploring historical migrant letter corpora\n\nThis project examines experiences of migration contained in one of the largest historical collections of migrant letters existing today, those of Irish migrants in the US and Canada during the late 18th to early 20th century. At present, the corpus contains around 4,000 texts (more are being added). The letters have been digitised and basic contextual information has been captured, such as author/recipient name and location, date of letter, relationship between participants (sibling, parent, grandparent etc.) and sex of author/recipient. The letters were collected from two sources: 1) the Documenting Ireland: Parliament, People and Migration (DIPPAM) project, an online archive hosted by Queen’s University, Belfast and 2) Professor Kerby Miller’s private collection of over 5,000 Irish migrant letters, housed at the University of Missouri. \nOver the past decade there has been a growing interest in migrant letters and ‘histories from below’. Social and cultural historians have overwhelmingly used qualitative methods to study both the complex social processes of migration (such as push/pull factors and the role of institutions and communities) and the conditions and daily lives of the migrants themselves (e.g. Erickson, Miller, Kamphoefner et al.). On the other hand, linguists have used mainly quantitative methods to explore language change and variation, identity construction and pragmatic features of correspondence (e.g. Elspa?, McLelland, Nurmi and Palander-Collin). There is, in other words, a separation of work between on the one hand the detailed study of individual migrant families and their trajectories, political contexts and socio-cultural worlds, and, on the other hand, the study of how linguistic features differ and evolve. The digitisation of migrant letters, as will be discussed, makes it increasingly possible for these disciplines to work together in new and creative ways. However, finding the best means to interrogate and compare large digitised letter collections is still a challenge. Issues such as lack of punctuation and non-standard spellings are problematic, and finding ways of scaling up micro-studies (for e.g., identifying topics and themes in the discourse of one letter collection and then doing the same across the whole corpus) present technical challenges that cannot easily be addressed using corpus tools and techniques alone. \nIn this presentation I will outline some of the work that has already been done with the migrant letter corpus (mainly by corpus linguists, historians and political scientists). I will then discuss some of the problems we have faced when working with the corpus and where I see a need for specialists in NLP, data mining and computer science in order to fully explore this type of data. \n\nhttps://www.csc.liv.ac.uk/research/seminars/abstract.php?id=1109
LOCATION:
END:VEVENT
END:VCALENDAR