Over the past few months, our team have been laying the groundwork for our research into the Relaciones �ұ�Dz���á�ھ������� and getting to grips with our source material. Here’s a sneak-peak of our exponentially expanding GIS place-name layers:��

The sheer size and non-standardised format of the Relaciones has meant that studying these documents has previously relied on close-reading of the texts, limiting the scope of research. Approaching this study from an interdisciplinary perspective offers us the chance to engage with innovative computational methodologies to create new opportunities for the exploration and study of these historic documents, improving accessibility and broadening the scope for research.��

Some of the key problems we aim to address include: the capability of computational methods in dealing with multilingual corpora, the ambiguous nature of many place-names mentioned within the Relaciones, and the general inaccessibility of historical texts as large and complex as this.��

We will be tackling these problems collaboratively as an interdisciplinary team, ensuring that our research contributes to the advancements of each of our fields of study. Each team brings their own expertise to the project, and by working collaboratively we are better equipped to tackle the problems posed by large historical source materials such as the Relaciones �ұ�Dz���á�ھ�������.��

One of the key challenges we face with the Relaciones �ұ�Dz���á�ھ������� is that of linguistics. This multilingual corpus features a combination of Spanish and a number of indigenous languages (predominantly Nahuatl) throughout. The excerpt below demonstrates one of the linguistic issues we face in dealing with these historic documents. “Hun 4at” and “Oxi 4ahol” are the indigenous names in Mayan Quiche for two volcanoes referenced in the ����������ó�� de Santiago ���پ��ٱ�á��.��

With Natural Language Processing systems usually being trained with modern news text, they would be unable to recognise and tag words in an indigenous language such as Quiche, especially with the unfamiliar usage of a numerical character in a place-name. Computational methods for the analysis of language are continually improving, though their use in the analysis of historical texts and non-English languages still presents many challenges. Our project aims to address these problems and improve methods for the analysis of complex historical documents such as the Relaciones �ұ�Dz���á�ھ�������.��

Feel free to leave your comments and/or . If you would like to read more about individual members of our team, please see our page.��

A considerable amount of our time on the project so far has been dedicated to GIS, and to the creation of maps which reflect geographic boundaries as they existed in sixteenth century Middle America. The work of and has been an invaluable source of information, providing detailed explorations of the historical geography of New Spain in the years following Spanish arrival.��

The maps Gerhard produced focused largely on the administrative boundaries imposed by the Spanish in the sixteenth century. We have used these maps as a starting point, creating GIS layers which reflect boundaries as depicted by Gerhard:����

��

Whilst some of these administrative units were built upon pre-existing and well-established indigenous systems of governance, they ultimately reflect a Spanish administrative geography, divided into dioceses, audiencias, and provincias, with identified place-names being the sites of Spanish alcaldia mayores and corregimientos.����

Of course, Middle America’s historical geography did not begin with the establishment of Spanish administrative units, and we aim to produce a more representative image of Middle America’s sixteenth century geography. The identification of historic place-names will be key to our understanding of��indigenous and historic geographies, and we are currently creating GIS layers which locate these place-names.����

The process for doing this has been straightforward but very time-consuming, as it could only be semi-automated. We began with the digitisation of various geographic indexes included in the works of Peter Gerhard, Mercedes de la Garza, René Acuña, Francisco del Paso y Troncoso and Alejandra Moreno Toscano. These indexes contain thousands of place-names referenced in the Relaciones �ұ�Dz���á�ھ�������. With these lists compiled, we then had the task of cross-referencing, or joining, these tables with existing location data, derived from a number of sources including historical secondary sources and .����

Of course, changes over time and variations in spelling mean that joining these sets of data cannot rely solely on a computer recognising identical matches. Our UK Research Associate, Dr Raquel Liceras-Garrido in collaboration with our Mexico Research Associate Mariana Favila-Vázquez, has been spearheading the painstaking task of locating the thousands of place-names for which there was no match with existing geographic data. Already a monumental undertaking, the process is further complicated by the fact that there are often numerous names for the same place (and these numerous names often have numerous alternative spellings!) For example, throughout the historical record, present-day Ixtacamaxtitlán in Puebla state is known variously as San Francisco Iztaquimaxtitlan, S Francisco Iztaquimaxtitlan, Istac-ymachtitlan, Estacquimestitlan, Itztaquimitztitlan and Castilblanco.����

Place names are often repeated across America (don’t get us started on San Juan) and there are numerous cases where it is not possible to determine the exact location of a place-name. The expertise of our colleagues in Mexico, Mariana Favila Vázquez and Dr Diego Jimenez-Badillo, has been invaluable in the disambiguation of hundreds of these place-names. For place-names which we have been unable to locate, our Mexican team have undertaken historical research in order to assign them coordinates. For those place-names which have managed to evade all our investigations (for now!), our Mexico team have so far been able to at least determine the region in which they lie.��

This research will also prove invaluable in our experiments conducted in collaboration with the Portuguese team on automatic multilingual Named entity Recognition and place-name linguistic and geographic disambiguation, which is the next stage in our research. ��