Comparing and Combining Early Modern Irish and Scottish Land Records: New Transkribus and Natural Language Processing Approaches

比较和结合早期现代爱尔兰和苏格兰土地记录旨在将负责爱尔兰契约登记处和苏格兰Sasines登记处的存储库与学者，数字人文专家和用户团体一起探索如何使这些记录更容易访问和理解。苏格兰国家档案局（NRS）和爱尔兰财产登记局（PRAI）分别是负责Sasines和爱尔兰契约的档案机构。这些是欧洲早期现代有关土地所有权、转让和财政利用的最全面的记录之一。它们提供了理想的基础，围绕它促进爱尔兰和英国的两个国际重要档案馆之间的新的跨学科合作，在各自国家的高等教育部门，并与公民和公共利益相关者的参与。苏格兰记录涵盖1617年至今，爱尔兰契约1708年至今。它们为两个不同的早期现代北方欧洲社会之间的长期直接可比性提供了一个独特的机会。土地是两个社会的主要政治、经济、社会和文化基础。记录反映了这一中心地位。他们提供了详细的见解：土地所有权的模式;宗教和政治少数群体的边缘化;爱尔兰和苏格兰的国际和全球商业;妇女的待遇;土地经济的运作;以及长期的趋势，在地方命名，语言和土地使用。这些记录对学术和地方历史学家、法律的学者、地理学家、地名和家族史专家、历史语言学家和古文字学家具有多方面的意义。它们也是对文化和遗产提供者具有明显的国家和国际意义的主要社会和公民资源。虽然Deeds和Sasines是具有国际重要意义的历史和文化遗产资源，但以目前的形式，它们基本上无法进入。NRS将其记录数字化，并提供在线访问，效果非常好。但是，由于记录的法律的作用，搜索能力仍然受到限制，因为这种作用是以了解财产和买卖双方为前提的。与此同时，PRAI已于2019年6月成立了一个数字化战略咨询小组，并计划对整个转录纪念册系列进行数字化。随着《爱尔兰契约》的早期版本很快以数字形式出现，该项目将利用一系列研讨会在感兴趣的利益攸关方之间建立合作伙伴关系。它旨在汇集、联系和联合收割机来自爱尔兰和苏格兰的专门知识、知识和资源。通过这种方式，它将以英国-爱尔兰数字人文科学合作呼吁所设想的方式提供实质性和增强的好处。这四个研讨会将于2020年5月至2021年4月在都柏林、爱丁堡和格拉斯哥举行。他们将：开发和测试一个两阶段的过程：首先，确认Transkribus的文本识别技术（HTR）的能力，以提供一致的，高准确性（90%-92%匹配）来自两个不同配置的多个创作记录的翻译，其中包含五种语言：英语、盖尔语、盖尔语、拉丁语和苏格兰语;其次，与TCD的ADAPT研究中心团队合作，网络参与者将致力于构思，开发和测试自然语言处理（NLP）识别文本元素、实体识别和提取技术的能力，以便于标记用户组识别的人、地点和其他相关实体。在网络结束时，参与者将确定未来多机构、跨学科赠款申请的时间表，该申请将通过围绕实体提取的小规模电子数据交换开始数字桥梁建设。

Comparing and Combining Early Modern Irish and Scottish Land Records aims to bring the repositories responsible for the Irish Registry of Deeds and Scottish Register of Sasines together with academics, Digital Humanities specialists and user groups to explore how these records can be made more accessible and understandable. The National Records of Scotland (NRS) and Property Registration Authority of Ireland (PRAI) are the archival institutions responsible for the Sasines and the Irish Deeds respectively. These are among Europe's most comprehensive early modern records relating to the ownership, transfer and financial utilization of land. They offer the ideal basis around which to foster new interdisciplinary collaboration between two internationally significant archives in Ireland and the UK, the HE sector in the respective countries, and engagement with civic and public stakeholders. The Scottish records cover 1617 to the present and the Irish Deeds 1708 to the present. They offer a unique opportunity for direct comparability over a long duration between two different early modern Northern European societies. Land was the primary political, economic, social and cultural foundation of both societies. The records reflect this centrality. They provide detailed insights into: patterns of land ownership; the marginalisation of religious and political minorities; Ireland and Scotland's international and global commerce; the treatment of women; the functioning of the landed economy; as well as long term trends in place naming, language and land use. The records are of multifaceted relevance to academic and local historians, legal scholars, geographers, place name and family history specialists, historical linguists and paleographers. They are also a major social and civic resource of obvious national and international significance for cultural and heritage providers. Although the Deeds and Sasines are internationally significant historical and cultural heritage resources, they are largely inaccessible in their current form. The NRS has digitized its records and provides online access to excellent effect. But search capacities remain limited by the legal function of the records which presuppose knowledge of the property and buyer/seller. Meanwhile, PRAI has established a digitisation strategy advisory group in June 2019 and intends proceeding with digitization of the entire volume series of transcribed memorials. With the early years of the Irish Deeds soon to become available in digital form, the project will use a series of workshops to build partnership working between interested stakeholders. It aims to pool, inter-relate and combine expertise, knowledge and resources from Ireland and Scotland. In this way it will provide substantial and enhanced benefits in precisely the way envisaged by the UK-Ireland Collaboration in the Digital Humanities call. The four workshops will be held between May 2020 and April 2021 in Dublin, Edinburgh and Glasgow. They will: Develop and test a two stage process that:Firstly, confirms the capacity of Transkribus's text recognition technology (HTR) to deliver consistent, high accuracy (90%-92% match) transcriptions from two differently configured, multiple authored records containing five languages: English; Gaelic; Gaeilge; Latin; and Scots;Secondly, working collaboratively with TCD's ADAPT Research Center team, the network participants will aim to conceive, develop and test the ability of Natural Language Processing [NLP] to identify textual elements, entity identification and extraction techniques to facilitate the mark up of people, places and other relevance entities identified by user groups. By the end of the network participants will have identified an timescale for a future multi-institutional, trans-disciplinary grant application which will begin digital bridge building through small-scale electronic data exchange (EDI) centred around entity extraction.