Irish English: a book of invasions

In ‘Murdering the language’ Moya Cannon imagines Ireland as a shore washed over by human tides. Each invasion added fresh layers to landscape, community and language, until:

Like the medieval Leabhar Gabhála or ‘Book of Invasions’ – a mythical history of the arrival of various peoples in Ireland – the textual record is a depository of earlier Irish English (IrE). It is no complete record, for most language use is never recorded and most texts are lost; nor is it an exact record of the spoken vernacular – the written word never perfectly matches the spoken – but a representation of speech at some degree of remove. Yet, analysed with care, texts of all types can reveal linguistic developments, and sometimes preserve evidence of spoken vernaculars, too.

Short introductions to the history of IrE are available in Filppula (Reference Filppula1999: 4–11) and Amador-Moreno (Reference Amador-Moreno2010: 16–30), while Hickey (Reference Hickey2007) provides a comprehensive survey. Corrigan (Reference Corrigan2010) deals with Northern Ireland, and a companion volume for the Republic is expected in the future.

English arrived in Ireland with soldiers and settlers accompanying Anglo-Norman lords in 1169. They settled along the east coast, establishing a colony that persisted through the Late Middle Ages in and around early towns. But English led a precarious existence in Ireland for the next five centuries. The Anglo-Norman aristocracy became Irish in culture, customs and language, and the English colony itself was so Gaelicised by the fourteenth century that its rulers enacted laws enforcing the use of English among people of English descent (cf. Crowley, Reference Crowley2000: 14–17, 20–3).

The language's status changed after the plantations of the late sixteenth and seventeenth centuries – one of the largest migrations in early modern Europe – brought new British settlers. Numbers settling in Ireland equalled the combined British migrations to North America and the Caribbean (Canny, Reference Canny and Canny1994). Military victory over the Irish established the settler community's undisputed hegemony after 1690. By then, the British were about a quarter of Ireland's population (Fitzgerald and Lambkin, Reference Fitzgerald and Lambkin2008: n.p., Figure 1), but political and economic power was firmly in the hands of this British minority's leaders.

Like the medieval colonists, these ‘New English’ settled mainly in the east. Two British concentrations emerged: a northern region centred on Belfast Lough, and a southern one in lowland Leinster, with Dublin as its hub (Fitzgerald and Lambkin, Reference Fitzgerald and Lambkin2008: n.p., Map 2). Fitzgerald's (Reference Fitzgerald1984) language maps, based on nineteenth-century censuses, show these regions of heaviest British colonisation were almost wholly English-speaking by 1800.

By that date, English was spreading rapidly among the native Irish. Fitzgerald (Reference Fitzgerald1984) maps the retreat of Irish into the west; the obverse of this process was the spread of English. By 1900, Ireland was practically a monolingual, English-speaking territory, and it has remained so despite the Republic's official policy of Irish revival. Still, Irish remains a presence as an official language in the Republic; and it has recently been recognised as a minority language in Northern Ireland, too, as part of the 1998 Belfast Agreement, which also granted recognition to Ulster Scots in both parts of Ireland.Footnote ¹

The book of invasions that is IrE contains chapters on:

• • dialect contact between vernacular varieties of English and Scots, brought to Ireland in the twelfth and then again from the mid-sixteenth century;

• • language contact with Irish throughout its history; and

• • language shift, as the native Irish opted for English between about 1750 and 1900.

Medieval IrE was the earliest overseas English, while the IrE that evolved after the early modern colonisation was contemporaneous with North American varieties. Both were outcomes of processes of dialect contact, involving mixing and levelling, as well as borrowing and transfer. Among the major issues in scholarship on IrE are the relative contributions of the Irish substrate, retentions from Early Modern English, Scots influence in the north, and the effects of language universals, especially in relation to features that distinguish IrE from other varieties.

Documentation

The best place to start researching any aspect of IrE is A Source Book for Irish English (Hickey, Reference Hickey2002). This comprehensive bibliography (with CD-ROM) lists nearly 3500 books and articles, with concise critical annotations of many items. It also offers a historical introduction and survey of research (Hickey, Reference Hickey2002: 1–55). This invaluable record is regularly updated on the compiler's Irish English Resource Centre website (www.uni-due.de/IERC/).

Textual evidence for medieval IrE is scanty. McIntosh and Samuels (Reference McIntosh and Samuels1968) listed 45 texts and summarised their linguistic features. They concluded that medieval IrE was fairly homogeneous, showing signs of levelling between varieties from the English south-west and south-west Midlands (McIntosh and Samuels, Reference McIntosh and Samuels1968: 8–9).

Some of the most important medieval texts – the Kildare poems and extracts from the Loscombe manuscript – are in the Corpus of Irish English (Hickey, Reference Hickey2003). They have been used for studies charting features of medieval IrE and their survival into the early modern era and beyond (e.g., Hickey, Reference Hickey2005: 194–202; Reference Hickey2007: 54–66), including such salient features of Southern IrE as /d/ and /t/ for /ð/ and /θ/. There are other texts in printed collections or recent editions (e.g., Hogan, Reference Hogan1927: 78–85; Dolan, Reference Dolan and Deane1991; Lucas, Reference Lucas1995). Given the small number of texts in medieval IrE, it is surprising that there has to date been no comprehensive collection.

Fragments of rural vernacular Englishes in Fingal, north of Dublin, and the south-eastern baronies of Forth and Bargy, provide further evidence of continuity with medieval IrE. Eighteenth- and nineteenth-century texts from Forth and Bargy are available in Dolan and Ó Muirithe (Reference Dolan and Ó Muirithe1996 [1979]), which also has a modern glossary from the area. Again, these and texts representing Fingallian are included in Hickey (Reference Hickey2003). Hickey (Reference Hickey2007: 66–84) examines these texts in his historical survey, relating them to older medieval material and later IrE.

Evidence is more plentiful from the early modern period onwards. Bliss (Reference Bliss1979) is a collection of 27 texts dating from 1600 to 1740, with documentation of the texts' and authors' backgrounds, and discussion of linguistic features. This anthology was severely – and unfairly – criticised by reviewers, who queried the linguistic authenticity of the texts. Later assessments have been kinder: Bliss' texts have contributed to interesting findings on the after-perfect (McCafferty, Reference McCafferty2004a); and Kelly (Reference Kelly and Tristram2001) found the IrE in one of the earliest texts more authentic than critics allowed, showing clear, valid evidence of lexical, phonological and syntactic transfer from Irish.

Scots in Ireland has been studied in personal letters, church records, and other documents. Robinson (Reference Robinson1989) examined an early Scots letter from Ulster dated 1571, while Montgomery (Reference Montgomery1992) looked at the rapid attrition of Scots in the correspondence of one family over just twenty years (1609–31), and Montgomery and Gregg (Reference Montgomery, Gregg and Jones1997: 584–93) examined Scots features in documents of the seventeenth and eighteenth centuries. Montgomery and Gregg (Reference Montgomery, Gregg and Jones1997: 593–604) also looked at literary Ulster Scots from the eighteenth and nineteenth centuries. And Robinson has used many of these sources for his Ulster Scots grammar (Reference Robinson1997). However, despite growing interest since the emergence of the Ulster Scots language movement in the 1990s, and official recognition after the Belfast Agreement, there has been no attempt to compile a historical linguistic corpus. That said, Ferguson's (Reference Ferguson2008) mainly literary anthology includes illustrative texts from 1603 to the present, indicating the range of material that might be included in such a corpus.

An important source for IrE in the late eighteenth century is Thomas Sheridan's Rhetorical Grammar of the English Language (1781), which offers the Irish advice on pronunciation. Sheridan's insights into late eighteenth-century IrE were overlooked until recently used as documentation of earlier IrE phonology (e.g., Hickey, Reference Hickey2005: 178–88, Reference Hickey2007: 346). Hickey (Reference Hickey and Hickey2010) locates Sheridan in relation to literary sources and the material for Fingallian, Forth and Bargy, and Ulster Scots.

Thanks to comments by Sheridan and other eighteenth-century prescriptivists, there is useful evidence of IrE deviations from prescribed norms. The exhaustive survey of grammars by Sundby, Bjørge and Haugland (Reference Sundby, Bjørge and Haugland1991) is worth trawling through for condemnations of Irish rule-breaking, which often equates to dialect. And one later normative account of Belfast ‘provincialisms’ (Patterson, Reference Patterson1860) has helped document the vintage of certain phonological features of Northern IrE.

The nineteenth century saw the tentative beginnings of IrE dialectology (e.g., Hume, Reference Hume1878; Joyce, Reference Joyce1988 [1910]).Footnote ² And in the early twentieth century, philologists provided general, largely literary-based accounts (Hayden and Hartog, Reference Hayden and Hartog1909; van Hamel, Reference van Hamel1912; Hogan, Reference Hogan1927). Later in the century, dialectologists and linguists began serious work on IrE. Henry's detailed account of a Roscommon dialect (1957), and his substantial pilot study for a never completed all-Ireland dialect survey (1958), laid important foundations for future developments.

Robert Gregg conducted the only major survey to date of Ulster Scots in the 1950s and 1960s (e.g., 1985). And Brendan Adams produced numerous papers on Northern IrE (e.g., Adams, Reference Adams1964, Reference Adams, Barry and Tilling1986). Data for a Tape-Recorded Survey of Hiberno-English Speech (TRS) was recorded in the 1970s and 1980s (Barry, Reference Barry1981, Reference Barry, Bailey and Görlach1982). This material is now available digitally (Hickey, Reference Hickey2004), along with the recordings for the Sound Atlas of Irish English and data from the Survey of Irish English Usage, both resulting from fieldwork done from the mid-1990s to 2002.

A number of major studies of urban varieties have also been conducted. Chief among these are the sociolinguistic studies of Belfast by J. Milroy (Reference Milroy1981) and L. Milroy (Reference Milroy1987 [1981]). Harris (Reference Harris1985) used data from the Belfast surveys and the TRS to study the phonology of Northern IrE. Kingsmore (Reference Kingsmore1995) is a sociolinguistic survey of Coleraine, and (London)Derry English is the subject of McCafferty (2001). For Southern IrE, the main work in this tradition is Hickey (Reference Hickey2005), which contains a variationist investigation of Dublin English.

Internal history

Our knowledge of the internal (linguistic) history of IrE comes largely from case studies focused on particular linguistic features, the usage of certain writers, or limited periods. Few studies take a long-term diachronic perspective, tracing the development of features over time. Published work that does so (e.g., Bliss, Reference Bliss1979; Sullivan, Reference Sullivan1980; Hickey, Reference Hickey2003: 14–27; McCafferty, Reference McCafferty2004a) is based on literary data, which is sometimes regarded with suspicion, since literary representations were often used for anti-Irish parody.

A diachronic study that takes a long-term perspective is Sullivan's (Reference Sullivan1980) work on IrE literary dialect, which traced the use of phonological and syntactic features in dramas from the seventeenth century to the twentieth. Lexico-phonological transfers predominated in the earliest texts, but declined after 1700 and generally disappeared by 1800 (Sullivan, Reference Sullivan1980: 198–200). Syntactic features took over as bearers of Irishness: these generally did not appear until about 1700, and were infrequent in the eighteenth century, but showed increased frequency after 1800 (Sullivan, Reference Sullivan1980: 200–5). This pattern correlates well with Thomason's model of language contact (Reference Thomason2001: 68–74).

There is a tradition of studies documenting literary uses of IrE. One classic is Taniguchi (Reference Taniguchi1972 [1956]), a study drawing on works by many authors, but particularly J. M. Synge, who often portrayed an IrE heavily marked by transfer from Irish. Amador-Moreno (Reference Amador-Moreno2006) studied the language of the bilingual proletarian novelist Patrick MacGill. In articles on William Carleton, McCafferty studies his use of the be after Ving construction (2005), and examines his handling of features from various regions (Reference McCafferty, Amador-Moreno and Nunes2009). Hickey (Reference Hickey2005: 166–77) surveys the language of Dublin writers like Dion Boucicault and Sean O'Casey. And an innovation within this tradition is Walshe's (Reference Walshe2009) investigation of IrE in films.

The greater range of text types that are becoming available for linguistic study permit the kind of triangulation that offers the only practical way of empirically resolving issues of authenticity. For the moment, it is worth noting that, where empirical results from literary data and other sources are directly compared, the literary texts are shown to be accurate (e.g., Sullivan, Reference Sullivan1980; Kelly, Reference Kelly and Tristram2001; Hickey, Reference Hickey2005: 166–77; McCafferty, Reference McCafferty2005).

Empirical studies

Investigating the origins of IrE has been a major strand of research. IrE emerged in a territory colonised by English-speakers. Hence, contact with Irish, as well as dialect contact, and the influence of vernacular British Englishes have been major themes. Some early research attempted to trace the origins of IrE traits to Elizabethan English (e.g., Braidwood, 1964), while early dialect surveys looked at both Early Modern English and Irish inputs (Henry, Reference Henry1957, Reference Henry1958). However, Gregg (Reference Gregg1985) was primarily interested in the survival of Scots in parts of Ulster characterised by the prevalence of Scottish settlement, Presbyterian religion, and Scottish surnames.

An important branch of research on northern varieties is the ‘Janus hypothesis’ (cf. Montgomery and Robinson, Reference Montgomery, Robinson, Sture Ureland and Clarkson1996), which takes up Gregg's orientation and insists on the Scottish origins of Ulster Scots/Northern IrE features also found in parts of the United States. Unfortunately, such work downplays the fact that many ‘Scots’ features were shared historically by lowland Scotland and the English north, and even the north Midlands. Since the latter regions supplied most of the English settlers in Ulster, and many of the English in the rest of Ireland too (MacCarthy-Morrogh, Reference MacCarthy-Morrogh1986; Canny, Reference Canny and Canny1994), it is hardly surprising that IrE today aligns with the ‘northern’ region identified by, e.g., Kortmann (Reference Kortmann, Kortmann and Upton2008: 491). This shared ‘northernness’ is marginalised in recent research concerned with Ulster Scots (e.g., Montgomery, Reference Montgomery2006).

But the search for Ulster Scots roots of US dialect features has also stimulated research using personal letters, a text type containing language closer to colloquial or vernacular usage than literary texts. Michael Montgomery (e.g., Reference Montgomery1989, Reference Montgomery1995) was the pioneer here. However, the general point that historical studies of IrE tend to have a narrow focus also applies to this research.

Thanks largely to Montgomery, the best studied feature of IrE historically is the Northern Subject Rule (NSR), a concord system that in its usual variable form permits either –s or zero forms of verbs with third-person plural subjects under certain conditions. Thus, The men goes or The men go are both grammatical, because the subject is a noun. But –s forms are prohibited with adjacent third plural pronouns, so that only They go is allowed. But with non-adjacent third plural pronouns – They always goes – the –s form is grammatical.

There have been numerous letter-based studies of the NSR in Ulster (e.g., Montgomery, Reference Montgomery1992, Reference Montgomery1997), and a fascinating tale trails the NSR from Scotland to Ulster, then North America and beyond (Montgomery, Reference Montgomery1999). However, the NSR has been shown to be more robust in nineteenth-century IrE in English-settled parts of Ulster than in areas settled from Scotland (McCafferty, Reference McCafferty2003). And it was also stronger in Southern IrE where Scots did not settle (McCafferty, Reference McCafferty2004b). Finally, comparing the use of the NSR by William Carleton (Kallen, Reference Kallen1991) and letters by a man who was a close contemporary of his, from the same district and similar social background, McCafferty (Reference McCafferty2005) found the two closely matched in their use of the NSR. The general northernness of IrE helps explain all these findings.

Another widely studied feature is the be after Ving construction (also ‘after-perfect’ or ‘hot-news perfect’), a clear transfer from Irish. Today, it refers to situations in the immediate or recent past, so that I'm after having a cup of coffee means ‘I've just had a cup of coffee.’ Older literary uses with future time reference were used to argue against the authenticity of literary representations. However, Bliss (Reference Bliss1979: 300) thought it better to seek explanations for such usages, and several scholars have looked into this.

Filppula (Reference Filppula1999: 99–107) suggested that an ‘old’ after-construction was replaced by a ‘new’ one by the early nineteenth century, and a later diachronic survey of usage from 1670 to 2000 (McCafferty, Reference McCafferty2004a) suggests this is correct. The construction has had categorically perfective uses since about 1850, but was mainly used with future reference until 1800. McCafferty (Reference McCafferty2005) studied the usage of the bilingual author William Carleton (1796–1869), who used this construction equally frequently with future and perfective meanings. Carleton was also placed diachronically in relation to other authors: those born before 1800 tended to use be after Ving with future meanings; those born after 1800 showed a preference for exclusively perfective meaning. Ó Corráin (Reference Ó Corráin and Tristram2006) found evidence from sixteenth- and seventeenth-century texts in Irish that the source construction for the after-perfect was similarly used with both future and perfective meanings, and Hickey (Reference Hickey2007: 202) cited evidence for both uses in trial records involving Irish people in the Old Bailey Online (www.oldbaileyonline.org).

Another of Sullivan's syntactic features is the IrE habitual aspect, as in He does be here in the mornings and He bes there every day, the former more typical of Southern IrE, the latter of northern varieties. Sullivan's analysis showed this was in use since the late eighteenth century, but became more frequent in late nineteenth- and twentieth-century dramas (Sullivan, Reference Sullivan1980: 203). Interestingly, his literary data attest the structure earlier than studies of Ulster emigrant letters, which only document the habitual from the mid-nineteenth century (Montgomery and Kirk, Reference Montgomery and Kirk1996).

The medial object perfect (MOP) in, e.g., I have my dinner eaten, was once regarded as a transfer from Irish. It then came to be seen as a retention of an older English perfective construction (e.g., Harris, Reference Harris, Trudgill and Chambers1991; Filppula, Reference Filppula1999). Lately, Pietsch (Reference Pietsch2009) has studied this construction in the Hamburg Corpus of Irish English and made a strong case for the MOP being an IrE ‘reinvention’. In this account, the MOP first fell into disuse, but reappeared in nineteenth-century IrE, modelled on a parallel Irish perfective construction.

The above amounts to a good deal of research on the internal history of IrE. Nonetheless, the typical historical investigation is a case study of a particular writer, genre or text type, a particular period, or a single construction. Diachronic studies using data permitting a long-term perspective are in a small minority. However, enough has been said about a variety of data sources for IrE for the observant reader to object that it must be possible to collate material from various sources for viable diachronic study. This is true to some extent already, and new research resources are currently being developed.

To test the usefulness of their new corpus, CORIECOR, the Corpus of Irish English Correspondence, McCafferty and Amador-Moreno (forthcoming) have studied two features for which it has been claimed that IrE influence caused changes in North American English that have since encroached on British English. These are the use of will rather than shall with first-person subjects, and the increase in the use of the progressive in Late Modern English.

That the Irish used will instead of shall with first-person subjects was a favourite complaint of normative grammarians (cf. Beal, Reference Beal2004: 96–7). Shall is virtually non-existent in present-day IrE (Kallen and Kirk, Reference Kallen, Kirk, Kirk and Ó Baoill2001: 71–3; Hickey, Reference Hickey2007: 179; Corrigan, Reference Corrigan2010: 64–5). However, our investigation (McCafferty and Amador-Moreno, forthcoming) revealed that first-person shall was the predominant form in late eighteenth-century IrE, as it was in American and Canadian English (Dollinger, Reference Dollinger2008). But in both IrE and Canadian English, will increased rapidly in the nineteenth century, becoming the majority variant by 1850. In IrE, it became virtually categorical only after the 1880s. IrE was thus unlikely to have driven the shift to will in North American Englishes.

A study of the progressive based on CORIECOR data showed the progressive to be far more frequent in late eighteenth-century IrE than in the comparable data of the Corpus of Late Eighteenth-Century Prose (Denison and van Bergen, Reference Denison and van Bergen2007). Its use increased more rapidly in nineteenth-century IrE than in any other variety for which data are available: Australian English (Fritz, Reference Fritz2007), British English (Smitterberg, Reference Smitterberg2005), or British and American English (Arnaud, Reference Arnaud1998). In this case, IrE might have contributed to the spread of the progressive where Irish immigrants were a considerable population group in the nineteenth century.

The future of Irish English's past

The future of IrE's past will be corpus-based. A number of corpora of IrE have been compiled, and some have been mentioned above. Among resources already available, being developed, or planned, are:

• • Conversational data from the Tape-Recorded Survey of Hiberno-English Speech (TRS) is available as the Northern Ireland Transcribed Corpus of Speech (Kirk, Reference Kirk2004). The TRS sampled three generations in its localities. Also available on DVD with Hickey (Reference Hickey2004).

• • ICE-Ireland (Kallen and Kirk, Reference Kallen and Kirk2008): a corpus of late-twentieth-century IrE with 1 million words of spoken and written data from Northern Ireland and the Republic.

• • SPICE-Ireland (Kallen and Kirk, Reference Kallen and Kirk2009): the spoken part of ICE-Ireland with pragmatic and discourse annotation, and prosodic transcriptions of one-third of the texts.

• • The Corpus of Irish English (Hickey, Reference Hickey2003) contains medieval poetry and dramas from the sixteenth century to the twentieth, as well as some prose texts.

• • The Hamburg Corpus of Irish English (http://www.uni-hamburg.de/sfb538/projekth5_e.html) is a collection of mainly nineteenth-century letters comprising about 600,000 words.

• • The Corpus of Irish English Correspondence or CORIECOR (McCafferty and Amador-Moreno, in preparation): personal letters from the seventeenth century to the twentieth. Currently, it comprises about 3 million words.

• • The 1641 Depositions Online (www.1641.tcd.ie) contain tribunal evidence given all over Ireland and may provide data on seventeenth-century IrE. The database is being adapted for linguistic research by a team at the University of Aberdeen and Trinity College Dublin.

• • There are also plans to digitalise the Irish Folklore Archive at University College Dublin for linguistic and other research purposes.

These resources span the history of IrE from the medieval period to the present. In the long term, it would be useful to combine many of these, and add further material, to create a diachronic megacorpus of IrE, but that is a task for the future. Meanwhile, what we have already provides a solid basis for adding more detailed empirical diachronic studies to our book of invasions.