(1) Overview
Repository location
The dataset ‘Irish and the Ordnance Survey Letters – Dialects, Accent Features, and Levels of Irish’ is hosted on Zenodo – 10.5281/zenodo.21384418.
Context
This dataset emerged out of research for a conference paper at the iClave 13 Language, im/mobility, belonging conference at the University of Lausanne, Switzerland, held between 29th June and 2nd July 2026, ‘Irish Dialectology and the Ordnance Survey of Ireland: Towards Standardisation before An Caighdeán Oifigiúil’. Previous works on the Ordnance Survey have focused on how the Survey was carried out (Andrews, 1976; Doherty, 2004) or there has been a specific focus on folklore and ethnography, particularly how information pertaining to these two subjects were depicted in the Letters and Memoirs (Ó Cadhla, 2007, 2011; McDonough, 2019). To my knowledge, there has yet to be a study of the Irish language as recorded in the Letters. This dataset addresses this gap.
(2) Method
Steps
In order to compile this dataset, the Ordnance Survey Letters were close read for the required information. The original letters are kept in the Royal Irish Academy in Dublin. These have been digitised and are available on the Ask About Ireland website (https://www.askaboutireland.ie/reading-room/digital-book-collection/digital-books-by-subject/ordnance-survey-of-irelan/), however, the photos of the Letters are often not legible. For reasons of speed and access to resources, the Michael O’Flanagan transcriptions of the Letters (O’Flanagan, 1927–1934) were used, which were published in the 1920s and 1930s. They were transcribed by a number of typists working under O’Flanagan. These are available on the Digital Repository of Ireland, where they were deposited by the OS 200 Ireland Mapped project and are available at the following link: https://dri.ie/os200/spotlight/os200/catalog?f%5Breadonly_collection_ssim%5D%5B%5D=OS+Letters&per_page=10 (See also Griffith et al. 2025). The digitised volumes in PDF from the DRI were downloaded and read closely to identify the relevant information. Once a passage had been identified as being relevant, it was transcribed manually into a spreadsheet with the following information: which county; which county was actually being referred to; which volume in the case of multi-volume counties; townland/parish name in English; townland/parish name in Irish; co-ordinates; date of letter; author name; recipient name; name of any local informant; information about Irish (dialects, accents, level of Irish spoken or not in a locality). The townland and parish names, along with the co-ordinates, were added after the other information had been entered.
Relevancy was determined by the following criteria. Firstly, that the Irish or the discussion of Irish, was not simply about a translation of a word into or from Irish. As the purpose of the Topographical Section of the Ordnance Survey of Ireland was to collect Irish place names and to provide translations or transliterations, this would have not provided any new information, as such is available on the Logainm website. Secondly, that for discussions about the level of Irish extant in a locality, only what was written in the Letters was recorded. For example, when it came to descriptions about the Irish spoken as being, for example, ‘good’, only this was recorded without judgment as to whether this meant a high number of speakers or a high level of quality of Irish. All mentions of Irish being spoken or not in a locality were recorded in the dataset.
Thirdly, when it came to dialectal and accent features, only those which mentioned a difference either between that of another locality or one which was considered interesting by the author of the Letter were recorded. The reason for this is that it is commonly understood that every region in Ireland has an accent and a dialect. By including only those which were considered noteworthy by the author of the Letter, this allows for the tracing of said pronunciation and dialectal word. For example, the difference in stresss patterns in the Letter dated 18th June 1836 from Mohill, Co. Leitrim, which recounts the distinction between stress patterns between two parishes only ca. 20 miles apart.
Sampling strategy
Letters exist for 29 of the 32 counties of Ireland; they do not exist for the counties of Cork, Tyrone, or Antrim. The Letters were close read and information for Irish in a townland or parish was sought out The reason for the Topographical Section’s existence was for the scholars employed by them to collect place names from local informants in Ireland, no information about the translation of Irish place names in to English, or vice versa was included. An exception to this was if it included a dialectal variant which was of interest to the scholar collecting it, as this demonstrated the geographical reach of a particular word. Information about the pronunciation of local place names was also not included, as the Topographical section was also created to sensitively transliterate Irish place names into English. Notes on accents where it demonstrates the intelligibility of different dialects and their pronunciation between the scholars and the local community have been included. As the Letters are based on the subjective sampling of the authors of the Letters within their own remit, naturally this sampling strategy was repeated in the dataset, as they recorded only which they found of interest. The aim of the dataset is to create a portrait of Irish as spoken in the 1830s and 1840s and where it was still strong or if the locality had shifted to English. To make sure that this aim was at the focus of the dataset, no mention of secondary Irish source material was included, even when it was quoted in Irish. This included a discussion of Mac Firbis’s Brehon law glossary in the Londonderry Letters. Where there was discussion about if Irish had been spoken in a particular locality in the previous decades or if there was mention of literacy in Irish which was no longer extant, these mentions were included.
Quality control
As one of the tasks of the Topographical Section was to standardise place names, the ones they give in the Letters vary in spelling and form. To make the data findable and reproduceable, the place names of townlands and parishes from the Logaimn website (logaimn.ie) were included, which are partly based on the work of the Ordnance Survey. Their versions for both the English and the Irish forms were used. In order to make the data findable and reproduceable, co-ordinates to highlight the specific townland were included. This is to specify where there are several localities of the same name, as well as for future use involving mapping. When a townland was referred to in the Letters, the co-ordinates of the townland were given. The same was applied to parishes and towns/cities.
(3) Dataset Description
Repository name
Zenodo.
Object name
Irish and the Ordnance Survey Letters – Dialects, Accent Features, and Levels of Irish.
Format names and versions
The data exists in two files – Excel and CSV, along with a description of the dataset and a methodology in a Word document.
Creation dates
The data were collected between 2026-06-08 and 2026-07-15.
Dataset creators
The data were collected and the dataset created by Ciaran McDonough.
Language
The data are in Irish and English, with a small amount in Latin.
License
The data have been deposited under Creative Commons Attribution 4.0.
Publication date
The dataset was published on Zenodo on 2026-07-15.
(4) Reuse Potential
The dataset has great reuse potential for those interested in historical sociolinguistics, dialectology, and nineteenth-century Irish history. The dataset has been curated to allow for further investigation by researchers; to form the basis of further datasets involving Irish linguistics – either involving data from other time periods or using other data from the nineteenth century. Appropriately, for data emerging from the Ordnance Survey of Ireland, the data could be used to map features of Irish in the nineteenth century. Because the data, and arranged deliberately so in my dataset, are arranged on a townland and parish level, it is possible to map dialectal differences on a national or provincial level; to trace the development of Anglicisation, as reported in the Letters; or to map phonetic features – both in terms of contemporaneous features and those from later dialect surveys of Irish. The world that the Letters portray disappeared less than a decade later as the Great Famine swept across the country, so, in collaboration with the other Ordnance Survey outputs (maps, antiquarian drawings, and Memoirs), it is possible to digitally recreate a world which would vanish.
In terms of limitations, I acknowledge that I am not a linguist, therefore some more specific linguistic features are probably missing from my dataset which would be desirable. However, these can be added to when the data are reused by those more linguistically inclined.
Acknowledgements
I acknowledge the Library of the Royal Irish Academy, Dublin who hold the Ordnance Survey Letters and who aided in the digitisation of them for the OS 200 Ireland Mapped project.
Data Accessibility Statement
The dataset which is described in this data paper is openly available in Zenodo at 10.5281/zenodo.21384418.
AI Declaration
No Generative AI was used in the drafting of this data paper or in the collection of data.