The Print Cultures of the Celtic Languages, 1700–1900

Provided by the author(s) and NUI Galway in accordance with publisher policies. Please cite the published version when available. Title Author(s) The print cultures of the Celtic languages, 1700–1900 Ó Ciosáin, Niall Publication Date 2015-05-01 Publication Information Ciosáin, Niall Ó. (2013). The Print Cultures of the Celtic Languages, 1700–1900. Cultural and Social History, 10(3), 347-367. doi: 10.2752/147800413X13661166397184 Publisher Taylor & Francis (Routledge) Link to publisher's version http://dx.doi.org/10.2752/147800413X13661166397184 Item record http://hdl.handle.net/10379/6274 DOI http://dx.doi.org/10.2752/147800413X13661166397184 Downloaded 2020-06-11T13:53:58Z Some rights reserved. For more information, please see the item record link above. [This paper was later published in Cultural and Social History X (2013), 347-67] THE PRINT CULTURES OF THE CELTIC LANGUAGES, 1700-1900 [Revised version, April 2012] Niall Ó Ciosáin School of Humanities, National University of Ireland, Galway [email protected] 1 Abstract: While the cultural trajectories of the Celtic language communities have some broad similarities in the long term, their histories in the medium term were quite different. This article approaches this issue through a comparative analysis of the print cultures of Welsh, Scottish Gaelic, Breton and Irish in the eighteenth and nineteenth centuries. The approach is both quantitative and qualitative, surveying total production in the four languages as well as looking at the presence and absence of different genres in the different languages. It also examines diaspora publishing in America and Australia. The different patterns are explained primarily in terms of the nature and extent of institutional church support for publishing in those languages. Keywords: print culture, literacy, language shift, publishing, reading. The Celtic languages and their communities have had a common experience over the past few centuries. In the face of English and French, the languages of two of the most powerful, centralized and expansionary European states, Irish, Gaelic, Welsh and Breton have all been in decline and retreat over the longer term. They were the subject of repressive policies by the central states, policies which were most pronounced in the sixteenth and seventeenth centuries; they were subject to less coercive but arguably more powerful structural forces towards linguistic uniformity, such as the integration of national markets and the growth of cash economies, pressures which were more dominant in the eighteenth and particularly the nineteenth centuries.1 At the same time, however, there were significant contrasts between the patterns of change of the different languages, and these do not correspond to differences in the coercive or persuasive forces involved. During the nineteenth century, for example, the number of Irish speakers declined rapidly and catastrophically, whereas the number of Welsh speakers increased substantially, perhaps even doubling (although given the growth in the overall Welsh population, the proportion of Welsh speakers did not, and neither did the number of monoglots). The number of speakers of Breton probably remained static or increased slightly; at any rate the borders between the Breton-speaking regions and the French-speaking regions were very close in 1900 to where they had been a century before, while the population had grown. The experience of Gaelic in Scotland was one of decline, though less rapid than that of Irish.2 For simplicity, I shall use ‘Gaelic’ to refer to Scottish Gaelic and ‘Irish’ to refer to Irish Gaelic. 2 Garrett Fitzgerald, ‘Estimates for Baronies of Minimum Level of Irish-Speaking amongst Successive Decennial Cohorts, 1771-81 to 1861-71’ Proceedings of the Royal Irish Academy 84 C (1984), pp. 117–155; Dot Jones, Statistical Evidence Relating to the Welsh Language, 18011911 (Cardiff, 1998); Brinley Thomas, ‘A cauldron of rebirth: population and the Welsh 1 2 These contrasts pose problems for the overall explanation of language shift outlined above. Wales was integrated politically with England as early as the midsixteenth century, and participated fully in the British industrial revolution. Ireland was integrated far later and less completely than Wales, and substantially deindustrialized in the nineteenth century. Nevertheless, up to the early twentieth century Welsh survived and even prospered, whereas Irish collapsed.3 Within an economic and political explanation of language shift, prominence is frequently given to literacy as a primary force. Individuals and communities had greater contact over time with documents, handwritten and printed, which necessitated literacy. This literacy was overwhelmingly achieved and practiced in the official language, and the usual mechanism through which literacy was acquired, the school, was also one of the principal agents of language shift. A straightforward equation of literacy and language shift is problematic, however, since it ignores the possibility of literacy in the unofficial language. There were in fact reading communities of varying sizes in all of the four main Celtic languages at all times. Moreover, literacy is not language-specific, but can be transferable between languages (where the languages share the same writing system). Reading ability that has been acquired in an official language can be applied to the local or home language without great difficulty, and this was very common in areas where nonofficial languages were spoken. A Welsh clergyman observed that in nineteenthcentury Brittany, 'the children, in learning to read French, acquire at the same time the ability to read their own language. You are aware that, in this indirect way, multitudes of children in Wales learn to read their own language'. This same process was envisaged in the preface to one edition of the most frequently published book in Irish: ‘every peasant who speaks Irish and reads English can master the work in its present form’.4 Such a conception of literacy as transferable, not language-specific, has a double implication. It means that, in practice, among bilingual populations, the size of the potential reading public is the same for both languages. This in turn underlines the importance of the supply of written material in determining whether or not a large reading public developed in a particular language. For the most part, that material will be printed, since manuscripts are too costly and too slowly produced to achieve mass circulation. language in the nineteenth century’, Welsh Historical Review 13 (1987), pp.418-37; Charles Withers, Gaelic in Scotland 1698-1981: The Geographical History of a Language (Edinburgh, 1984); Fanch Broudic, La Pratique du Breton de l’Ancien Régime à nos jours (Rennes, 1995), pp.33-58, 414; 3 Thomas, ‘A cauldron of rebirth'. 4 Nancy H. Hornberger, ‘Continua of Biliteracy’, Review of Educational Research 59 (1989), pp.271-296; Thomas Phillips, Wales: the language, social condition, moral character, and religious opinions of the people, considered in their relation to education (1849), p.568; Timothy O’Sullivan [Tadhg Gaelach Ó Súilleabháin], Pious Miscellany (Dublin, 1858); 3 Quantitative comparisons The Celtic languages developed varying levels of print culture, strongest in Welsh, weakest in Irish. These differences can be approached initially in terms of simple quantity, and a table of book and pamphlet titles published during the nineteenth century presents this in the starkest terms [Fig. 1]. The standard figure given for Welsh is that presented to a Royal Commission in 1896 by the bibliographer Charles Ashton, who estimated that nearly 8,500 items had been published. As Ashton himself pointed out, this was far from a complete total, and the real figure is perhaps nine or even ten thousand. Titles in Breton, as computed in the 1970s by Yves le Berre, add up to about 1,200, while the Scottish Gaelic Union Catalogue lists almost 1,000 items of four pages or more. The lowest by far is Irish, with less than 150. This figure is calculated from library catalogues and other sources, and it is revealing of the marginalization of print culture in the Irish language that there is no comprehensive modern bibliography of printed books produced in the eighteenth and nineteenth centuries.5 These figures of course represent approximations, in that they have been compiled in different ways and do not take account of factors such as survival rates. Nevertheless, even doubling the figures for particular languages does not fundamentally alter the overall contrasts between them, which are orders of magnitude rather than simple quantities. Report of the Royal Commission on Land in Wales and Monmouthshire HC 1896 XXXIV, p.901; Philip Henry Jones: ‘A golden age reappraised: Welsh-language publishing in the nineteenth century’, in Peter Isaac and Barry McKay (eds), Images and Texts: Their Production and Distribution in the Eighteenth and Nineteenth Centuries (Winchester, 1997), pp.121-41; Yves le Berre, La Littérature de la Langue Bretonne: Livres et brochures entre 1790 et 1918 (3 Vols., Brest, 1998); Michel Lagrée, ‘La littérature religieuse dans la production bretonne imprimée: aspects quantitatifs’, in Michel Lagrée (ed.), Les Parlers de la Foi: Religion et langues régionales (Rennes, 1995), pp.85-94; Mary Ferguson and Ann Matheson, Scottish Gaelic Union Catalogue: A List of Books Printed in Scottish Gaelic from 1567 to 1973 (Edinburgh, 1984); E.R. Dix and Séamus Ó Casaide, A list of all the Books, Pamphlets, etc. printed wholly or partly in Irish… to 1820 (Dublin 1905); Risteárd de Hae and Brighid Ní Dhonnchadha, Clár litridheacht na nua-Gaeilge 1850-1936 (3 vols., Dublin, 1938-40). 5 4 Fig.1. Printed production in the four Celtic languages, 1800-1900 These figures refer to the nineteenth century, but the pattern they record had been established well before then. The second half of the eighteenth century shows remarkably similar ratios, although the amounts are of course smaller [Fig. 2]. About 1,900 titles appeared in Welsh, with the average rising from 10 per annum in 1730s to 40 in the 1770s.6 In Gaelic, the figure is 59. This included 19 editions of the Psalms, a fundamental text of highland culture and the basis of a tradition of communal psalm singing, which had first appeared in 1659. The New Testament of 1767, by contrast, was very slow to sell its first edition of 10,000 copies. This period also saw the first publication of the most successful non-biblical text in Gaelic, Dugald Buchanan’s Laoidhe Spioradail, first published, like the New Testament in 1767, but which sold rapidly, going through 9 editions by 1800.7 There is no precise figure for Breton for this period. However, a guess can be made from some surviving documents. An inventory from the year 1777 for the firm of Blot, the sole printer in Quimper, lists more than twenty different works in 6 Geraint H Jenkins, The Foundations of Modern Wales 1642-1780 (Oxford, 1987), 409; Geraint H Jenkins, ‘The cultural uses of the Welsh language 1660-1800’, in Geraint H Jenkins (ed.), The Welsh Language before the Industrial Revolution (Cardiff, 1997), p.371; all books in Welsh up to 1820 are listed in Eiluned Rees, Libri Walliae: A Catalogue of Welsh Books and Books Printed in Wales 1546-1820 (Aberystwyth, 1987). 7 An Ceud Chaogad do Shalmaibh Dhaibhidh (Glasgow, 1659); Tiomnadh Nuadh ar Tighearna (Edinburgh, 1767); Dugald Buchanan, Comh-chruinneacha do Laoidhibh Spioradail (Edinburgh, 1767) 5 Breton, most in substantial print runs.8 Given that there were printers working in Breton in Morlaix and Brest in the same period, and given that the 1777 inventory represents production at only one point, an overall figure of 80 or 100 seems reasonable. Finally, there were some 20 editions of works in Irish, of which seven were editions of Sixteen Sermons in an Easy and Familiar Stile by James Gallagher, Catholic bishop of Raphoe, first published in Dublin in 1736 and the most frequently printed text in Irish in the eighteenth century. Fig.2: Printed production 1750-1900 These are remarkable differences, not simply of quantities, but of orders of magnitude. In fact, the contrasts could be said to be even greater than these figures suggest, since they do not take into account the relative sizes of the language communities. In 1800, there were perhaps two and a half million speakers of Irish, almost a million of Breton, half a million of Welsh and perhaps 300,000 of Gaelic. Adjusting for population therefore reinforces the dominance of Welsh printing, enhances the volume of Gaelic printing, and makes the paucity of print in Irish even more striking. Qualititative contrasts Simple quantitative comparison also fails to convey the different degrees of variety of production within the print culture of each language. This variety was in terms both of format and content, of printed forms and of genres of text. In most of Jean-Louis Le Floc’h and Gwennolé Le Menn, ‘L’Imprimerie-librairie Blot à Quimper en 1777’, Mémoires de la Société Historique et Archaeologique de Bretagne 62 (1985), pp.157-78. 8 6 these areas, the qualitative relationships between the different languages are the same as the quantitative relationships. A few aspects can be considered as representative. The relative strength of the periodical press in the different languages in the nineteenth century is one example. There were dozens of newspapers and periodicals in Welsh, some lasting for decades. The Methodist Y Drysorfa appeared continuously from 1830, while the Baptist Seren Gomer began as a weekly in 1814, was relaunched as a fortnightly in 1818, and was published monthly from 1820 throughout the nineteenth century. At the other end of the spectrum, a Welsh version of Punch, Y Punch Cymraeg, was published between 1858 and 1861.9 Breton had six periodicals, of which five were religious, notably the Catholic church’s Feiz ha Breiz, first published in 1865. In Gaelic, there were about a dozen, predominantly religious and moral in content and short-lived. The best-known were produced by Norman MacLeod, a minister of the Church of Scotland who wrote as ‘Caraid nan Gaidheal’, who established An Teachdaire Gaelach (1829-31), which contained a miscellany of religious, historical and other material, and Cuairtear nan Gleann (1840-3), which contained a good deal of material relating to emigration. In Irish, however, only one periodical for Irish speakers was published. This was An Fior Eirionnach, which appeared in the early 1860s and lasted for seven issues.10 Similar proportions apply to the printing of some of the commonest printed genres of the period. The best-selling secular printed text in all European countries was the annual almanac. Almanacs in Welsh were printed continuously from the 1680s onwards, and from the early eighteenth century there were competing almanacs from different printers. By the late nineteenth century, the market for Welsh-language almanacs was very substantial indeed, with competing almanacs being produced in print runs of tens of thousands. In Breton, almanacs were produced from at least the early nineteenth century by printers such as Alexandre Lédan in Morlaix in print runs of two thousand or so, while almanacs in Gaelic were published from the 1870s onwards. In Irish, however, only one almanac seems to have been produced, in 1724, and that was part of an experiment in Protestant evangelical printing rather than as a commercial proposition.11 Huw Walters, ‘The Welsh language and the periodical press’, in Geraint Jenkins (ed.), The Welsh Language and its Social Domains 1801-1911 (Cardiff, 2000), pp.349-78; for Punch, see Peter Lord, Words with Pictures: Images of Wales and Welsh Images in the Popular Press, 16401860 (Aberystwyth, 1995), pp.145-46. 10 Le Berre, Littérature, pp.158-9, 176; Sheila Kidd, ‘Norman MacLeod’ in John Koch (ed.), Celtic Culture: a historical encyclopaedia (Santa Barbara, 2006), pp.1223-4; Kidd, ‘Tormod MacLeòid: Àrd-Chonsal nan Gàidheal?’ in Sheila Kidd (ed.), Glasgow: Baile Mòr nan Gàidheal / City of the Gaels (Glasgow 2007), pp.107–29; Donald E. Meek, ‘The pulpit and the pen: clergy, orality and print in the Scottish Gaelic world’, Adam Fox and Daniel Woolf (ed.) The spoken word: Oral culture in Britain, 1500-1850, (Manchester, 2002), pp.84-118, 106-09. 11 Hans-Jurgen Lusebrink, ‘Traduire l’almanach populaire: essai de typologie et mise en perspective socio-culturelle’ in Lusebrink (ed), Les Lectures du peuple en Europe et dans les Amériques du XVIIe au XXe siècle, (Paris, 2003), pp.145-155; Geraint H. Jenkins, Literature, 9 7 The matter of another widespread cheap genre, the ballad, is a little complicated by the variety of formats in which ballads were printed, but once again Welsh has the longest and most substantial history of production, followed by Breton. Welsh ballads in the eighteenth century were produced in 8-page booklets, with 2-4 songs in each booklet. About 760 of these booklets are known. In the nineteenth century, the format was more frequently a 4-page booklet, and a recent bibliography lists 8,000 items, printed by 350 printers in nearly a hundred towns. The typical Breton format was the single sheet, and the standard bibliography lists over a thousand items, mainly from the nineteenth and early twentieth centuries. These were both religious, as in the classic ‘Cantiques’, and secular, usually referred to as ‘Gwerziou’. The Gwerziou frequently referred to current events such as wars and natural disasters.12 In Irish, there are a few dozen, most printed in the early nineteenth century.13 Similar proportions of printed production were replicated among the emigrant communities from Wales, Scotland and Ireland in North America and Australia. As early as 1720, a book of 120 pages in Welsh, Annerch ir Cymru (translated in 1727 as A Salutation to the Britains) by Ellis Pugh, was published in Philadelphia, and two others followed in the next twelve years.14 Major publishing in Welsh dates from the extensive migration that began in the late eighteenth century, from mining areas of Wales to mining areas of the United States such as Scranton in Pennsylvania and Utica in New York. Most of these settlers were monoglot Welsh speakers and formed a substantial market for publishing in Welsh there. This market was serviced partly by local printers and partly from Wales itself, with some Welsh publishers having agents in the US. In all, some 72 periodicals in Welsh were established between 1832 and the early twentieth century, although only one, Y Drych Americanaidd [The American Mirror], a weekly newspaper printed in Utica, appeared continuously from the mid-nineteenth into the twentieth century.15 Religion and Society in Wales, 1660-1730 (Cardiff, 1978), pp.36, 231-34; Gérard Bailloud, L’Imprimerie Lédan à Morlaix (1805-1880) et ses Impressions en Langue Bretonne (Saint-Brieuc, 1999), pp.44-5; Ferguson and Matheson, Scottish Gaelic Union Catalogue, nos 24, 980-91; Nicholas Williams, I bPrionta i Leabhar: Na Protastúin agus Cló na Gaeilge (Dublin, 1986), pp.125-27. 12 E. Wyn James, ‘Zulus and stone-breakers: a case study in Glamorgan ballad-sheet printing’, in Mary-Ann Constantine (ed.), Ballads in Wales (London, 1999), pp.41-48; Joseph Ollivier, Catalogue Bibliographique de la Chanson Populaire Bretonne sur Feuilles Volantes (Quimper, 1942); Daniel Giraudon, Chansons Populaires de Basse-Bretagne sur Feuilles Volantes (Morlaix, 1985). 13 ‘Alf Mac Lochlainn, ‘Broadside ballads in Irish’, Éigse 12 (1967-68), pp.115-22; Hugh Shields, ‘Nineteenth-century Irish song chapbooks and ballad sheets’, in Peter Fox (ed.), Treasures of the Library (Dublin, 1986), pp.197-204. 14 D.H.E. Roberts, ‘Welsh publishing in the United States of America’, in P. H. Jones, E. Rees (eds), A Nation and its Books: A History of the Book in Wales (Aberystwyth, 1998), pp.253-64. 15 Roberts, ‘Welsh publishing’; Deian Rhys Hopkin, ‘Welsh immigrants to the United States and their press, 1840-1930’, in Christiane Harzig and Dirk Hoerder (eds), The Press of Labor Migrants in Europe and North America, 1880s to 1930s (Bremen, 1985), pp.349-67; William D. Jones, Wales in America: Scranton and the Welsh, 1860-1920 (Cardiff, 1993). 8 Gaelic printing in North America was on a far smaller scale and was concentrated in Canada, to which there was substantial migration from the Highlands and Islands during the nineteenth century. Most of the books and periodicals in Gaelic were produced in Cape Breton and in Montreal, where editions of the two best-selling Gaelic books, the hymn collections of Dugald Buchanan and Peter Grant, were printed in 1836. The first of a series of periodicals was An Cuairtear Og Gaidhealach [The Young Gaelic Visitor], 13 monthly issues of which appeared in Nova Scotia in the early 1850s. The most successful of the periodicals was Mac-Talla [Echo], published in Cape Breton, weekly from 1892 to 1901 and fortnightly until 1904.16 Printing in Irish in North America, by contrast, was almost non-existent even though emigration from Ireland to North America was proportionately among the highest in Europe during the nineteenth century, and included many Irish speakers. An Irish-language paper, An Gaodhal [The Gael], began to be published in New York in the 1880s, but as its aim was revivalist, it was directed at those who wished to learn Irish as much as to those whose spoken language it was.17 The same phenomenon almost exactly can be observed on a far smaller scale in Australia, where a monthly magazine, Yr Australydd [The Australian] appeared in the mining town of Ballarat, Victoria, between 1866 and 1874, followed by essentially the same magazine published in Melbourne under the title Yr Ymwelydd [The Visitor] from 1874 to 1876. A Scottish Gaelic monthly had been published even earlier in Tasmania in 1857, and lasted for eleven issues. The first printing in Irish was again in a revivalist context. This was ‘Our Gaelic Column’ which appeared weekly or fortnightly in the Melbourne Advocate between 1901 and 1912.18 The similarities between the publication patterns in North America and in Australia are striking. The ratios of printed material are the same, as are the proportions of speakers, Welsh least, Irish greatest. While local circumstances in the new countries must have played some part, and the settlement patterns were probably more favourable to publishing among the Welsh who settled in relatively dense mining communities, it is clear that the patterns of publishing and reading were largely brought by the emigrants and reflected those in the countries of origin. Charles Dunn, The Highland Settler: a portrait of the Scottish Gael in Nova Scotia (Toronto, 1953), Ch. 6, ‘Gaelic in Print’, pp.174-190; J. L. Campbell, ‘Scottish Gaelic in Canada’, American Speech 11 (1936), pp.128-36; Kenneth E. Nilsen, ‘Some notes on pre-Mac-Talla Gaelic publishing in Nova Scotia’, in Colm Ó Baoill and Nancy R. McGuire (eds), Rannsachadh na Gàidhlig 2000 (Obar Dheathain, 2002), pp.127-40. 17 Fionnuala Uí Fhlannagáin, Mícheál Ó Lócháin agus An Gaodhal (Dublin, 1990). 18 Robert Tyler, ‘The Welsh Language Press in Colonial Victoria’, Victorian Historical Journal, vol. 80, 2009, pp.45-60; Val Noone, ‘‘Our Gaelic Column’, Melbourne 1901-1912’ in Brad & Kathryn Patterson (eds.), Ireland and the Irish Antipodes: one world or worlds apart? (Sydney, 2010), pp.162-174; Niall Ó Ciosáin, ‘Old languages in a new country: publishing and reading in the Celtic languages in nineteenth-century Australia’ The Australasian Journal of Irish Studies XI (2011), 58-72. 16 9 First-generation migrants from Wales, and to a lesser extent from Scotland, were used to reading in Welsh and Gaelic, and prospective publishers had textual models they could adopt. The Tasmanian Gaelic monthly was in fact called An Teachdaire Gaidhealach [The Gaelic Messenger], in explicit homage to Norman MacLeod’s first periodical three decades earlier in Scotland, while the many Welsh monthlies in North America echo the periodical press in Wales. One of the most remarkable features of these contrasts is the paucity of print in Irish, and almost everywhere one looks, this pattern is confirmed. There were, for example, Welsh-language books printed in Ireland but no Irish-language books were printed in Wales. The first edition of Gwaedd Ynghymru by the Puritan Morgan Llwyd was printed in Dublin in 1653, while some ten books in Welsh were printed there in the second half of the eighteenth century and a Welsh language almanac as late as 1805. Indeed more Welsh-language books were printed in Dublin during the 1740s and 1750s than Irish-language ones (Fig.3).19 Fig. 3: Gair i’r Methodist, Dublin 1751. As regards Irish and Gaelic, perhaps the most striking contrast is in the printing of collections of song by Jacobite and other poets. The pioneer in Scotland was Alasdair Mac Mhaighistir Alasdair, whose Aiseirighdhe na Sean Chanoin 19 Charles Parry, ‘From Manuscript to Print 2: printed books’, in R. Geraint Gruffydd (ed.), A Guide to Welsh Literature Vol.III: 1530-1700 (Cardiff, 1997), pp.263-76, 267; Rees, Libri Walliae nos. 128, 489, 1717, 2499, 3408, 3479, 4424, 4425, 4438-39, 4443, 4453, 4600, 4690, 4758, 4966, 5099, 5127-28 10 Albannaich; no, An nuadh oranaiche Gaidhealach [The Resurrection of the Old Scottish Language] was published in 1751. He was followed by Duncan Ban McIntyre, whose Orain Ghaidhealach [Gaelic Songs] was printed in 1768, 1790 and 1806. The 1790 edition has a 40-page list of subscribers. There were many other anthologies of poetry and song which collected work from poets since the sixteenth century. One of these, Patrick Turner’s Comhchruinneacha do dh’Orain Taghta, Ghaidhealach [Anthology of Selected Gaelic Songs] of 1813 has a subscription list of almost 1,700 names. In Ireland, by contrast, although the tradition of composition of poetry in Irish was at least as strong as in Gaelic, not a single eighteenth-century poet saw his work in print.20 As regards Irish and Breton, we can compare two popular or folk poets of the early nineteenth century. In some respects, they were remarkably similar. Anthony Raftery (1784-1835) and Iann Le Gwen (1774-1849) were both blind itinerant poets who frequented fairs and pilgrimages selling similar compositions about the news and events of the day. Both composed songs on the cholera epidemic of 1832, for example, and on contemporary shipwrecks. However, Le Gwen sold his songs as printed sheets, produced in Morlaix by Alexandre Lédan, himself a composer of songs in Breton. I know of no evidence that Raftery had any contact with print and printers.21 There were, moreover, printers in Brittany and Wales who specialized in books in Breton or Welsh, but there was no Irish printer in the eighteenth or nineteenth century who could be said to have specialized in books or other texts in Irish. Chronology of production We have established that the four languages had radically different experiences of print culture, both in quantity and in quality. What lies behind these differences? An explanation emerges if we examine another quantitative aspect of print, that is, the chronology of production in the different regions during the nineteenth century. In the second half of the century, a marked divergence is visible between the development of print in Irish on the one hand and print in the other three languages on the other. Niall Ó Ciosáin, ‘Popular song, readers and language: printed anthologies in Irish and Scottish Gaelic, 1780-1820’ in John Kirk, Michael Brown and Andrew Noble (eds.), Cultures of Radicalism in Britain and Ireland (forthcoming 2012). 21 Daniel Giraudon, ‘Chanteurs populaires et production imprimée au XIXe siècle en Basse Bretagne’, Mémoires de la Société Historique et Archaeologique de Bretagne 62 (1985), pp.41525; L. M. Cullen, ‘Poetry, Culture and Politics’, Studia Celtica Japonica 8 (1996) pp.1-26; Ciarán Ó Coigligh (ed.), Raiftearaí: amhráin agus dánta (Dublin 1987); Michel Lagrée, ‘Foi et langue en Bretagne et en Irlande au XIXe siècle’, in Yann Celton et al (eds.), Chrétientés de Basse-Bretagne et d'Ailleurs (Quimper, 1998), pp.275-281. 20 11 Fig. 4: Printing in Breton 1790-1919 Fig. 5: Printing in Gaelic 1700-1900 12 Fig. 6: Devotional works in Irish 1720-1900 Fig. 7: Welsh newspapers of more than five years duration in circulation, 1800-1900 As with fig.1, the numbers for Breton are those compiled by Le Berre, and those for Gaelic are based on the Scottish Gaelic Union Catalogue. The graph for Irish shows devotional works only, but as noted above, devotional books represent the vast 13 majority of production.22 The graph for Welsh is a proxy, since there is no full bibliography of publications after 1820. It shows the number of newspapers in circulation and illustrates a thriving print culture.23 Overall, these tables show printing in Welsh, Breton and Gaelic continuing to expand after 1850, and indeed achieving their greatest strength in that period, and at the same time the collapse of a print culture in Irish that was much slighter than the others to begin with. Moreover, as with total production, the graph understates the contrast between Irish and the others, because it doesn’t take account of print runs. The industrialization of book production after 1850, along with improvements in distribution, meant that these could be much greater then before. Some Welsh almanacs had print runs of between 40,000 and 70,000 copies in the years 1871 to 1905, for example, whereas it is unlikely that any book in Irish in the nineteenth century had a run of more than a thousand or two.24 What was it about Ireland or the Irish language that caused this chronological difference with the other three Celtic language areas? One possibility would be the Great Famine of 1845-50, which had a catastrophic effect on Irish-speaking areas, and the subsequent population decline which was most marked there. This would certainly be in contrast to Welsh, whose number of speakers doubled between 1850 and 1900 to 2 million, or Breton, whose numbers remained steady or even increased slightly in the same period. On the other hand, the experience of Gaelic was similar to that of Irish. The Highlands and Islands had a potato famine in 1846-7, and although it was on nothing like the scale of the Irish one, many Gaelic-speaking areas had dramatic losses in population, mainly due to emigration. The island of Barra lost a third of its population between 1841 and 1851, as did Mull. The Gaelicspeaking population was in decline thereafter, as was the Irish. This did not, however, result in a dramatic fall in printing in Gaelic, rather the contrary.25 Questions of population, therefore, do not explain the anomaly of Irish. Other considerations reinforce this conclusion. In the first place, the decline in printing in Irish is far more dramatic than that in population. In the second place, as far as print culture is concerned, the number of readers is more important than the number of speakers, and the second half of the nineteenth century was a time of 22 This graph is based on the useful but incomplete bibliography of devotional works printed in Irish in Malachy McKenna, ‘A textual history of The Spiritual Rose’, Clogher Record 14 (1991), pp.52-73, with additions of my own. 23 I have adapted this graph from Aled Jones, ‘The Welsh language and journalism’, in Jenkins, The Welsh Language and its Social Domains, pp.379-403, 381. Jones’ graph shows the number of Welsh-language newspapers launched in each decade, without an indication of how long they survived. The present graph shows the number of newspapers in existence in any decade, and is a more accurate reflection of the buoyancy of the sector. I took the figures from the same source as Jones, that is, Beti Jones, Newsplan: report of the Newsplan project in Wales (Aberystwyth, 1994). 24 Jones, ‘A golden age reappraised’, p.122. 25 T. M. Devine, The Great Highland Famine (Edinburgh, 1988); for late nineteenth-century figures, see the works in note 2. 14 rapidly increasing literacy in Irish-speaking areas. This literacy was of course achieved initially in English, and is usually thought of as an anglicising force. However, as noted above, literacy can be transferable between languages, and many, perhaps most, readers of Irish at this time had learnt their letters in English originally. This is clear from the use of anglicised spelling in many catechisms and single-sheet ballads in Irish as well as from the preface to the 1858 edition of O’Sullivan’s Pious Miscellany quoted earlier. There was, therefore, a readership in Irish-speaking areas which was potentially far greater than the printed production would suggest. It was not, however, provided with material by publishers and printers, whereas readers in the other language areas were, and there is no obvious socio-economic or political explanation for this. It may be more illuminating to turn the question around. Why was the later nineteenth century the high point of production and reading in the other three areas? The answer to this lies in the predominantly religious nature of print culture in all four languages. The fact that the bestsellers in Irish and Gaelic were hymn collections has already been noted, and much the same is true of Breton, while periodicals in Welsh, Breton and Gaelic were predominantly religious in content. In essence, what sustained print culture in Welsh, Breton and Gaelic in the later nineteenth century, and what was absent for Irish in that period, was support from institutional churches. Moreover, in earlier centuries, the development and nature of a print culture was largely the result of campaigns and interventions on the part of religious institutions. We can illustrate this by looking at the four regions in turn. The role of churches in the later nineteenth century In the later nineteenth century, there were significant elements within the institutional churches in Wales, Scotland and Brittany that sponsored the publication of reading material in their respective languages, and envisaged a far greater role for them in religious practice, public or domestic, than was the case with Irish. The clearest example is the Free Church of Scotland, founded in 1843 as a breakaway from the state church, the Church of Scotland. The majority of the Gaelic-speaking population joined the Free Church, and from the beginning it emphasised Gaelic and the reading of Gaelic as part of its identity. It published a magazine in Gaelic, An Fhanuis [The Witness], which by 1846 claimed a circulation of 5,000, and the death in 1847 of its first moderator, Thomas Chalmers, was marked by the publication of a long Gaelic lament composed by a Free Church Minister. In Brittany, most of the periodicals in Breton were published under diocesan auspices, and the church mobilised and revived the traditional genre of the 15 cantique, printing collections of these hymns for particular occasions such as missions or for general use in a specific diocese.26 The link between religious denominations and publishing is also clear in the case of the Welsh periodical press. The report of the 1896 commission quoted earlier discusses them in precisely these terms: There is no daily paper published in that language, but there is a good number of weekly ones, of which some are more or less closely identified with individual religious denominations. Such are, for instance, y Goleuad, 'the Luminary' connected with the Calvinist Methodists; y Tyst, 'the Witness' with the Independents; Seren Cymru "the Star of Wales" with the Baptists; y Hwyliedydd, 'the Sentinel' with the Wesleyans; and y Llan a'r Dywysogaeth 'the Church and the Principality' with the Church of England. The report makes the same observation with regard to monthlies, of which 'several... are issued in connexion with particular religious denominations.'27 It was not that the churches in Scotland, Wales and Brittany were promoting language revival or a full-blown linguistic nationalism. Nevertheless, they all displayed a pragmatic support for publishing as part of catechesis and evangelisation, and in all cases, influential elements within those churches went beyond such pragmatism and suggested a broad connection between vernacular languages and the identity of a particular church. In Wales and the Highlands of Scotland, the nonconformist denominations and the Free Church defined themselves partly by their use of Welsh and Gaelic, as against the landed elite who were more anglicised and belonged to the state church. The 1845 General Assembly of the Free Church, just two years after its foundation, was inaugurated by sermons in both English and Gaelic, had a daily Gaelic service, and saw the collection of the melodies of the Gaelic psalter which were subsequently published along with the texts.28 In Wales, the connection between religion, language and identity was consolidated by the ‘Blue Books’ controversy of 1847, in which a Royal Commission on education poured scorn on both the Welsh language and non-conformity, provoking an angry reaction throughout Wales. The following decades saw a major popular political mobilisation that took place through Welsh to a great extent and Select Committee on Sites for Churches (Scotland), 2nd Report, Minutes of evidence H.C. 1847 XIII, 119; Donnchadh Mac Gilleadhain, Cumha an Diadhair Urramaich Dr. Thomas Chalmers... (Glasgow, 1848); The Breton periodicals are listed in Le Berre, Littérature, 91-2. 27 Royal Commission on Land (see note 5), 653. 28 Proceedings of the General Assembly of the Free Church of Scotland held at Inverness, August 1845 (Edinburgh, 1846); T.L. Hately, Seann fhuin nan Salm mar tha iad air an seinn anns a' ghaeltachd mu thuath.... (Edinburgh, 1845). 26 16 which was partly directed at the religious disabilities of nonconformists and ultimately at the disestablishment of the state church.29 In Brittany, some influential elements in the Catholic church were by the later nineteenth century insisting on the close relationship between religion and language, encapsulated in the title of the most prominent of the Catholic periodicals, Feiz ha Breizh, 'Faith and Brittany', founded in 1865. This was part of a more general orientation in the French Catholic church, in which similar attitudes were emerging with regard to other regional languages. This can be illustrated by the case of the Marian apparitions that were such a prominent feature of nineteenth-century French Catholicism. At La Salette in 1846, and again at Lourdes in 1854, messages were delivered by the Virgin Mary in the local languages, Provençal and Béarnais respectively, and their language was urged by many within the church as proof of their authenticity. Satellite shrines of La Salette were soon established in Brittany, cantiques about it were published in Breton and were popular enough for others to be requested in the religious press.30 Moreover, the increasingly bitter conflict over education between church and state towards the end of the century also added to the tendency within the church to identify Breton with Catholicism, in opposition to the French-speaking secular central state.31 It should also be pointed out that the churches as institutions were not necessarily unanimous or uniform in their attitude towards the use of vernaculars. This is particularly the case with the Catholic church in Brittany, where for example a policy of not appointing natives of a diocese as bishops there meant that the vast majority of bishops in Brittany in the nineteenth century did not speak Breton. There were also powerful elements within the church, such as teaching orders, whose practice was to encourage assimilation to metropolitan French culture and who were dismissive of local languages, seeing them as doomed relics of the past. A wide range of opinion was expressed in clerical debates, ranging from suggestions that the survival of christianity was tied up with the survival of Breton on the one Gwyneth Tyson Roberts, The Language of the Blue Books: the perfect instrument of empire (Cardiff, 1998); Prys Morgan, ‘From long knives to blue books’ in R.R. Davies et al (eds.), Welsh Society and Nationhood; historical essays presented to Glanmor Williams (Cardiff, 1984) 199-215; 30 Michel Lagrée, ‘Langue céleste et langue régionale au XIXe siècle’, Annales de Bretagne et des Pays de l’Ouest 98 (1991), 121-29; René Merle, ‘L’Apparition de La Salette et les "patois"’, Lengas 31 (1992), 69-105; Michel Lagrée and Georges Prévost, ‘La Salette en Bretagne’, in François Angelier and Claude Langlois (eds), La Salette: apocalypse, pélerinage et littérature (Grenoble, 2000), 155-70; ‘Gwerz an itron-varia-ar-zalet', Journal Général de l'Imprimerie et de la Librairie, 106, Pt. 1 (1869), no. 5117; 'Eur C’hantik da Itron Varia ar Zalett’, Feiz ha Breiz Sept.-Oct 1905; for the Catholic church’s attitude to regional languages in France generally, see Gérard Cholvy, ‘Régionalisme et clergé catholique au XIXe siècle’, in Christian Gras and George Livet (eds), Régions et Régionalisme en France du XVIIe siècle à nos jours (Paris, 1977), 187-201; in Ireland, the principal Marian apparition took place in a partly Irish-speaking area in Mayo in 1879, and was, tellingly, silent. 31 Michel Lagrée, Religion et Cultures en Bretagne, 1850-1950 (Paris 1992); Caroline Ford, Creating the Nation in Provincial France: religion and political identity in Brittany (Princeton, 1993), 24-8 29 17 hand, to the view that Breton was unable to express complex philosophical or theological ideas on the other. The majority, however, favoured at the very least a pragmatic approach whereby the use of Breton was necessary to effective ministry, and consequently supported a continuous publication campaign in Breton.32 Support for publishing in the vernacular, and for its use in religious services, was far from implying approval of all of the rest of the culture of those languages. The Free Church was severely critical of traditional Gaelic sociability, of music, song, dance and storytelling, while the Catholic church likewise campaigned against many elements of Breton culture, including aspects of lay religious manifestations such as religious theatre and pilgrimage. Indeed campaigns of evangelisation and catechesis in the vernacular are largely predicated on such hostility and on a desire to reform belief and behaviour in the target communities. In the second half of the nineteenth century, therefore, publishing and printing in Breton, Welsh and Gaelic were all promoted and sustained by influential elements within the institutional churches to which the majority of their speakers belonged. Moreover, support for the language in this period came to form part of a broader anti-metropolitan mass politics. The case of Ireland appears initially to be similar. Almost all Irish speakers were Catholics, and the politics of the Catholic church in Ireland were similar to those of the churches in the other areas. It was, for example, in continuous conflict and negociation with the state over education and its control, as were the Catholic church in France and the nonconformist churches in Wales; its clergy were prominent in nationalist and agrarian political organisations and it sought disestablishment of the state church, like Methodists in Wales; and it ended up supporting the Liberal coalition in the Westminster Parliament, also like the Methodists and the Free Church. In stark contrast, however, it displayed little or no interest in Irish as part of its cultural politics, or even a pragmatic interest in printing in Irish for pastoral purposes. There was no Catholic periodical at all in Irish in the later nineteenth century, and very little religious publishing generally in Irish, despite the fact that in the decades after the Great Famine, the Catholic church had emerged as the dominant social and ideological force in Ireland and that Irish was still the language of up to a quarter of the population. Its campaign against unorthodox local religious belief and practice, which was similar in other respects to that in Highland Scotland and in Brittany, was conducted largely in English, to such an extent that one of the most influential interpretations of this process could suggest that new forms of For the range of opinion within the clergy and hierarchy, see Lagrée, Religion et Cultures, 221-44; for debates among seminarians at the end of the century, Christian Brule, ‘L’Académie bretonne au grand séminaire de Quimper’ in Lagrée (ed.), Les parlers de la foi 32 18 orthodox Catholicism acted as a substitute for the Irish language as a force for social cohesion.33 Conclusion: churches, print norms and reading publics The above discussion has focused on the second half of the nineteenth century, when church sponsorship of vernacular publishing was at its height. However, as we saw earlier, the same ratios of printed material were visible in earlier periods also, when that sponsorship was less obvious. Moreover, throughout the nineteenth century there were secular publications in all languages whose existence was not attributable to church or religious impulses. In Gaelic, there was a continuing production of verse or song, as well as the reprinting of older anthologies; in Brittany, Le Berre's figures show that about half of printed production in Breton can be described as secular; and Welsh had the most varied corpus of all, including a ten-volume encyclopaedia published between 1856 and 1879.34 How does an argument based on church sponsorship of publishing account for these? The answer is that in earlier periods, elements within institutional churches played a crucial role in creating print cultures in the different languages in the first place. They provided the basic texts, such as catechisms and hymns, along with the Bible in Welsh and Gaelic, and in that process they set the norms of orthography, typography and language for printed works that were followed by later publishers, and thereby created a rudimentary reading public in a language. This public then became the prospective audience for later publications, secular as well as religious. The relative quantities and strengths of the print cultures of the different languages are related to how early these institutional interventions took place - the late sixteenth and early seventeenth century for Welsh, the late seventeenth century for Breton and the later eighteenth century for Gaelic. In Welsh, the crucial text was the Bible. The New Testament was published by the Church of England in 1567 and the Old in 1588. These were large and expensive editions, however, designed for church use. Much more influential in setting a norm and forming a readership was the series of octavo editions of the same texts which appeared from the 1620s onwards and which were designed for domestic use.35 In the following century, this print norm was successfully exploited by Methodist and other non-conformist groups, who mobilised a major print campaign from the mid-eighteenth century onwards. In Breton, a print norm was largely created in the mid to late seventeenth century, during a series of intensive Jesuit missions to Breton-speaking parishes. These were accompanied by the Emmet Larkin, 'The Devotional Revolution in Ireland', American Historical Review 77 (1972), 625-652. 34 Le Berre, Littérature; Glanmor Williams, Religion, Language and Nationality in Wales (Cardiff, 1979), 142 35 Parry: ‘From Manuscript to Print’, p.266. 33 19 composition and distribution of printed devotional texts in Breton, such as catechisms and, in particular, canticles. These were mostly written by Julien Maunoir, the principal missionary in Brittany, who was not a native speaker and whose written Breton was heavily influenced by French, both in language and orthography. This style of Breton later became known as ‘priests’ Breton’ (‘brezhoneg beleg') and characterised much of the printed production thereafter.36 As in the nineteenth century, this was part of a wider pattern of engagement by the Catholic clergy with local languages for catechetical purposes during the CounterReformation. This engagement produced an identification of local languages and Catholicism against which the later policy of linguistic centralisation pursued by the Revolutionary regimes was a reaction.37 Finally, in Gaelic, the establishment of a printed norm is usually traced to the publication of the New Testament of 1767 and of the Old Testament in 1801, sponsored by the Scottish Society for Promoting Christian Knowledge (SSPCK). As in the later nineteenth century, the churches were not unanimous in their support of these initiatives, and they did not imply approval of the cultures of the different languages. In this earlier period, moreover, these publishing projects were not continuous, but were concentrated in particular periods. The established church in Wales did not continue its publishing campaign much beyond the early seventeenth century, and by the eighteenth was identified with the use of English, not least by the new Nonconformist denominations who were mobilising Welsh print. In Brittany, the Jesuit effort was concentrated between the 1640s and 1680s, falling off completely after 1700, while in the Highlands the SSPCK had existed for sixty years before the publication of the New Testament. As regards unanimity, the Jesuit missions were opposed by the secular clergy of the time, including by the bishops of the breton dioceses. The SSPCK in its earlier years aimed more at extirpating Gaelic than printing it, and its commissioning of the Gaelic Bible marked a major shift in policy. In that same period, however, the most frequently printed text in Gaelic was produced by the Church of Scotland. This was the translation of the psalms commissioned and published by the Synod of Argyll in 1659. As for the wider vernacular culture, the attitude of those clergy who used the local languages was if anything even more negative than in the later period. Maunoir was convinced that diabolic witchcraft was rampant in Brittany and was also reported by the antiquarian scholar Edward Lluyd in 1708 to have burnt every non- Maunoir formulated a simplified orthography in his Sacré College de Jésus (16 59), see Ropartz Hemon (ed.), Doctin on Christenien (Dublin, 1977) pp.viii-ix and Iwan Wmffre, Breton Orthographies and Dialects: The Twentieth-Century Orthography War in Brittany (Berne, 2007), pp.16-18. 37 David A. Bell, ' Lingua Populi, Lingua Dei: Language, Religion, and the Origins of French Revolutionary Nationalism', American Historical Review, 100 (1995), 1403-1437, part. 142531. 36 20 devotional manuscript in Breton that he could find. The distance between Maunoir and the culture of his audiences can be measured by his repeated comparisons of the Bretons with the Iroquois and Huron, among whom Jesuit missionaries were also active in this period.38 In all three languages, therefore, we can say that the existence of a reading public and of a popular print culture were the result of decisive interventions by institutional churches or by elements within those churches. Moreover, the resulting print cultures remained substantially or even largely religious. The bestsellers in Gaelic, as we saw, were the hymn collections of Buchanan and Grant, while the variety of religious publishing in Welsh is remarkable. The most successful and influential breton authors of the eighteenth century were two priests, Charles le Bris and Claude Guillaume de Marigo.39 Le Bris' book of hours was first published in 1712 and had 36 subsequent editions, as well as numerous abridgments. Marigo is best known for his collection of saints' lives, Buez ar Saent (1752), which became the most widely owned Breton book in the nineteenth century, each farmhouse supposedly possessing a copy.40 A continuing link with the institutional church can be seen in an edition of 1837, which contains an imprimatur from the bishop of Quimper dated the previous year. Like Maunoir, Le Bris and Marigo produced texts that were, like Maunoir's, strongly influenced by French, both in vocabulary and orthography. This form of print, suitable for an audience which, as noted earlier, read French as well as Breton, was the model for later religious writers.41 The success and durability of the print norms in Breton and in Welsh can be gauged by the hostile responses to later attempts at changing the orthography of the two languages in print. For Welsh, at the end of the eighteenth century, a new orthography was elaborated by William Owen Pugh, and in 1804 the British and Foreign Bible Society proposed to print their Bibles in this new form. By all accounts Pugh’s orthography was eccentric, but it is revealing that the principal objection made to the new Bible was that readers wouldn’t be able to understand it because it departed from established norms. The previous editor of the Society’s Bibles, John Roberts, went as far as to say that ‘Like the British Constitution, our Welsh orthography is already fixed and established’.42 In Breton, a new orthography, more Eric Lebel (ed.), Miracles et sabbats. Journal du Père Maunoir: missions en Bretagne, 16311650 (Paris, 1998); Martin Harney, Good Father in Brittany: the life of blessed Julien Maunoir (Boston, 1964), 72; Lluyd quoted in H.L. Humphreys: 'The Breton Language: its present position and historical background' in M. Ball and J. Fife (eds.): The Celtic Languages (London, 1993) 606-643, 619. 39 Yves le Gallo, Clergé, Religion et Société en Basse Bretagne: de la fin de l'ancien régime à 1840 (Paris, 1991), 442-4 40 Ford, Creating the Nation, 63 41 For Marigo and Le Bris as linguistic models for a later clerical writer, see the preface to M. Marrec, Doctrin ar Guir Gristen (Saint Brieuc, 1846) 42 R. Tudur Jones, ‘The church and the Welsh language in the nineteenth century’, in Jenkins, The Welsh Language and its Social Domains, pp.215-37, quote on p. 232. 38 21 differentiated from that of French, was elaborated by Le Gonidec in 1807 and was favoured by cultural nationalists such as La Villemarqué, who adopted it in his Barzaz Breizh. Its proposed use in religious periodicals in the diocese of Quimper from 1843, however, provoked a storm of protest from parish clergy. Their argument was a practical one, that readers would not understand a text which was not presented according to the established norms of Breton printing. 43 In Irish, however, in the absence of a concerted publishing effort by any church, no standard was established, and a very wide variety of printed forms persisted between the sixteenth and the nineteenth centuries. In the sixteenth and seventeenth centuries, Irish print was intended to be continuous with the manuscript tradition in appearance, using Gaelic founts which had to be specially manufactured, and even going to the extent of mimicking manuscript abbreviations, which are strictly speaking unnecessary in a printed book. This was true of both Protestant books and the Catholic works which were printed in Europe, mainly by the Franciscans in Louvain and Rome. (Fig.10) Beginning in the eighteenth century, however, Catholic printing for a popular readership, most of whom would in all probability have had little or no experience of manuscripts, was using roman letters and various degrees of orthographic simplification, assimilated, that is, to printing in English (Fig 11). In the early nineteenth century, by contrast, the Protestant evangelical societies which we have already mentioned reverted to the earlier style. In their educational campaigns, they recruited some of the remaining Irish-language scribes as translators and teachers, and the books produced under their influence resembled the manuscripts which they were used to producing and reading. Later language revivalists and cultural nationalists overwhelmingly used Gaelic founts, partly as a way of differentiating Irish from English, but in the process, like Le Gonidec in Brittany, produced books that were not comprehensible to a popular Irish-speaking readership. By the 1830s, therefore, there was a profusion of printed forms of Irish, from a Protestant manuscript style on the one hand to a variety of Catholic ones on the other, which used roman letters and a range of types of simplified orthography, and consequently no clear model for printers and readers to follow. The principal, perhaps the only, institution capable of elaborating and instituting such a model, the Catholic church, did not do so.44 43 Le Berre, Littérature, 36; Jean-Louis le Floc’h, ‘Controverses sur la langue bretonne dans le clergé finistérien au XIXe siècle’, Bulletin de la Sociéte Archaeologique de Finistere 114 (1984), pp.165-77. 44 There is a comprehensive discussion of Irish founts in Dermot McGuinne, Irish Type Design: A History of Printing Types in the Irish Character (Dublin, 1992). 22 Fig.10 A Catholic catechism, Rome 1676 23 Fig 11. Catholic catechism, Dublin 1802 To return to the wider question of language shift with which we began, there is no doubt that political and economic processes exerted a decisive influence on the fate of the Celtic languages in the longer term. However, the speed at which the shift took place varied enormously between the different language areas. This variation was due to the fact that the processes were mediated through widely different religious circumstances. Ultimately, it is the degree of support from institutional churches, through their construction of a print culture, which best explains the variety of paths taken by the Celtic language communities in the eighteenth and nineteenth centuries. 24