Using Intergenerational Transmission as a New Marker of Irish-Language Shift, 1851-1891

Nicholas Wolf

New York University

ACIS Mid-Atlantic, 2019

Unresolved Questions about 19th c. Language Shift

• How long did shift really take?
  • 3 generations, e.g. monoglot Ir, bilingual, monoglot E
  • "like the snow off the ditches": Charles McGlinchey (1861-1954) of Co. Donegal, The Last of the Name
  • All-island shift: 800+ years, still incomplete
  • Cos. Kerry, Galway, Donegal, e.g.: 300 years, incomplete
• Was rate of shift uniform, or did it differ by region?

Unresolved Questions about 19th c. Language Shift

• Did features of 19th c. Irish society accelerate or hinder shift?
  • Mass mortality? Mass emigration? Seán de Fréine, "An Gorta agus an Ghaeilge" (1995), regarding Famine as cause of shift: "Ach níl de bhunús ag an teoiric seo ach an méid seo, gur tharla na rudaí seo de bharr an Ghorta de bhrí gur tharla siad i ndiaidh an Ghorta. Is léir gur falsa an cineál réasúnaíochta seo." Famine not cause, but accelerant
  • Rural depletion (in favor of city growth)?

Unresolved Questions about 19th c. Language Shift

• What role did the centralization of administration (however incomplete) and of curriculum (very complete) of the education system impact language survival?
  • Maureen Wall (1969): jettison O'Connell-National Schools-Church as language-shift cause "catechism"

Unresolved Questions about 19th c. Language Shift

• How was shift experienced differently by gender?
  • Females as sociolinguistic innovators (drive shift), e.g. via "need to modernize"?
  • ... or males as sociolinguistic innovators , e.g. role as public actors?
• How did local community structure, occupational structure affect shift?
  • "Market-oriented" local economies and shift?
  • Farming/nonfarming divides?

Need for More Empirical, Less Anecdotal

• The marvels of the 1851-1891 Irish censuses and the question of language--a quick history of language measurement on censuses:
  • 1846, Belgium (nonstandard in how question asked)
  • 1860, Switzerland
  • 1880, Austria & Finland
  • 1881, India and Scotland
  • 1890, United States
  • 1891, Wales
• Micheál Ó Gliasáin, Ceist na Teanga sa Daoinaireamh/The Language Question in the Census of Population (1996)

Limitations of Irish Census, 1851-1891

• Undercounts
• Destruction of original returns in WWI
• Changing spatial granularity, limited variables of published tables (gender, location, bi/monolingualism)

Affordances of Irish Census, 1851-1891

Garret FitzGerald (1984, 2003), G.B. Adams (1975, 1979)

• Treat counts returns for each census year as a minimum count
• Focus on 1851-1881 censuses, and especially 1881, because of the move of language question out of footnote in 1881, helping to mitigate undercount
• Harmonize language data given in baronial spatial units with age cohort data given at PLU-Registrars/Dispensary District level (=education/health divide)

Affordances of Irish Census, 1851-1891

Garret FitzGerald (1984, 2003), G.B. Adams (1975, 1979)

• Spatial harmonization no small feat!

Affordances of Irish Census, 1851-1891

Garret FitzGerald (1984, 2003), G.B. Adams (1975, 1979)

• Spatial harmonization no small feat!

Disadvantages of Adams/FitzGerald Approach

• Lose longitudinal analysis ability
• Possibly requires too much apportionment of RDs across baronial borders
• Not much discussion of gender

New Approach: Parish-Barony and Age Structure Analysis

• Substitute Parish Age-Level information (available in literacy tables at less-granular age cohort sizes) for RD-based approach
• Potentially less need to apportion (Co. Clare, for example, has only one parish that crosses a baronial boundary)
• Lose some age granularity (0-19 years, 20-39, and 40 up) in harmonizing across census years, but potential gain in spatial precision.

New Approach: Pilot of Munster

• Double-key entry of all age/literacy tables (55% complete); double-key entry of language tables with help of QUB project from early 2000s.
• Focus on 1851/1871/1891 to enable 20-year age cohort longitudinal inspection (newborns to 19 year olds in 1851 become 20 to 39 year olds in 1871, etc.)

New Approach: Pilot of Munster

• Eliminate problem of undercounting by abandoning comparison of number of Irish speakers to number of English speakers
• Instead, compare the proportion of overall Irish-speaking population under 20 to the proportion of overall population under 20; disproportionately low shares of under-20 indicates weakening of Irish speaking
• Advantages: universally comparable (even outside of Ireland, e.g. Wales 1891)
• Key assumption: undercounts are unbiased with regard to age

New Approach: Pilot of Munster

New Approach: Pilot of Munster

1851 Females, Munster, Differences in Share of Population of Age 0-19 Cohorts

(Irish-speaking versus Population Overall)

New Approach: Pilot of Munster

1871 Females, Munster, Differences in Share of Population of Age 0-19 Cohorts

(Irish-speaking versus Population Overall)

New Approach: Pilot of Munster

1891 Females, Munster, Differences in Share of Population of Age 0-19 Cohorts

(Irish-speaking versus Population Overall)

Test 1: Correlation with Population Depletion, 1841-1851? (No...)

Blue = Females

Red = Males