Studying verbal change in Canadian English in real time: The Bank of Canadian EnglishLaurel J. BrintonUniversity of British ColumbiaFor a copy of this power point, go to: http://blogs.ubc.ca/englishlanguagestudies/
5th International Conference on Late Modern English – Bergamo, Italy
LMod
E-5
The history of post-colonial varieties• Post-colonial varieties of English have been well-studied, but their
histories remain understudied. • In large part, this is the result of the lack of readily accessible
(electronic) resources for historical study (except American English)• Canadian English (CanE) – have only CONTE-pC - 125,000 words, 1776–1849,
Ontario English, diaries, (semi-)official letters, local newspapers (see Dollinger 2008: 99ff.)
• American English (AmE) – COHA – 1810–2009 (400 million words) (Davies 2010–)• Australian English – 19th (COOEE) and 20th (AusCorp) centuries (c. 340,000 words)
(see Collins 2013)• New Zealand English – CENZE (c. 282,000 words) (see Hundt 2012)
• Beal, Fitzmaurice, and Hodson (2012: 205) argue that “[t]he importance of electronic corpora for the study of linguistic variation and change in [the LModE] period cannot be overstated … It is only by having access to large amounts of data and the tools with which to annotate and analyse these that we are able to see the patterns that do emerge”.
LMod
E-5
Scholarship on the history of CanE(see Dollinger [2012] for a recent overview)
• external history (settlement patterns)• e.g., Bailey (1982), Boberg (2010: 55-105), Chambers (1998, 2010)
• historical development of lexis• e.g. Avis et al. (1967), Story et al. (1999), Dollinger and Brinton
(2008)• historical phonology• e.g. Chambers (2006) on “Canadian raising”, also work on the
Canadian shift, low back merger, yod dropping• a wealth of apparent time studies, especially of urban varieties of CanE
(looking at a wide range of morphosyntactic features)• e.g. Tagliamonte (2006), Tagliamonte and D’Arcy (2007), Tagliamonte
(2013)• new dialect formation • e.g. Schneider (2007: 238–250) – transition from exonormativity (c.
1812–) through nativization (c. 1867–) to endonormativity (c. 1920–)
LM
odE-
5
Outline of the paper“Striking … is the absence of a diachronic, real-time perspective … in CanE” (Dollinger 2012: 1859)• Brinton and Fee (2001) – but not a corpus study• Dollinger (2008) – a study of modal auxiliaries in early Ontario
English based on CONTE-pC
This paper considers a “workaround” (Dollinger 2012: 1865) for the study of historical CanE using the Bank of Canadian English (BCE). In the paper, I will:• describe the nature of the BCE• present three case studies using the BCE
LMod
E-5
Bank of Canadian English (BCE)Lexicographic database (web-based database application used to collect citations) for the revision of A Dictionary of Canadianisms on Historical Principles (Avis et al. 1967; rpt. 1991)
Digitized first edition available online in open access: http://dchp.ca/DCHP-1/
LMod
E-5
Contents of the BCE• Approximately 30,000 citations from the 1st edition (DCHP-1)• 3,400 legacy citations (Strathy Unit, Queen’s University) (see
Dollinger 2006)• Newly collected citations for the 2nd edition (DCHP-2) for a total of
71,194 citations for 17,508 headwords (changing daily)• Current size of the BCE is 2,461,498 words
LMod
E-5
Length of citations
OED overall OED-2 OED Additions volumes
BCE overall0
50
100
150
200
250
# of characters per citation
Figure 1: Average citation lengths (OED figures based on Sheidlower 2011; see Brinton, Dollinger, and Fee 2012: section 3 )
LMod
E-5
Regional and temporal coverage• As far as our sources allowed, we have gathered a structured
data set (see Dollinger 2010) • For words arising after WWII, 10-year intervals were used, for
words older than WWII, 25-year intervals. • Data from as many provinces and territories as possible were
collected• For example, for gas bar, we have • 2000s – BC, AB, SK, MB, ON, QC, NS, PE, YT, NB• 1900s – BC, AB, SK, NB, ON• 1980s – AB, ON• 1970s – BC, ON• 1960s – ON
• This has yielded a hybrid between structured corpus and unstructured quotations database.
LMod
E-5
Temporal range: 1505-2013
Figure 2: Numbers of words per 20 year period in the Bank of Canadian English (1740-2013) (accessed 15 July 2013)
1500-
1739
1740-
1759
1760-
1779
1780-
1799
1800-
1819
1820-
1839
1840-
1859
1860-
1879
1880-
1899
1900-
1919
1920-
1939
1940-
1959
1960-
1979
1980-
1999
2000-
2013
0
100000
200000
300000
400000
500000
600000
700000
800000
LMod
E-5
Text types: BCE (pilot stage)
Figure 3: Proposed text type categories of the BCE (DCHP-1)
LMod
E-5
Text types: DCHP-2For DCHP-2 citations, we have used electronic resources (almost exclusively):• Canadian Newsstand: full-text access to nearly 300 Canadian newspapers
from all provinces and territories from 1977 to the present. • Canada’s Heritage from 1844 – The Globe and Mail • Toronto Star Pages of the Past (1894 to 2008)• Early Canadiana Online: digitized selection of the content from the
microfiche held by the Canadian Institute for Historical Microreproductions (up to 1920) – parliamentary records, novels, travel reports
• The Champlain Society Digital Collection: almost 50,000 printed pages) dealing with exploration and discovery from the 16th to 19th century
• British Columbia Historical Newspapers: collection 24 local newspapers, 1865-1924
• Peel’s Prairie Provinces: collection of 6500 books, 106 newspapers (from 1871), and other sources
• Various university student newspapers (e.g. The Ubyssey [1918-present])• and other smaller electronic sources
LMod
E-5
Limitations of DCHP-2 data collectionNewspaper archives, it is claimed, do not constitute a balanced or representative corpus (Atkins and Rundell 2008: 62)But one can argue that newspapers• include considerable quantities of recorded/represented speech.• can be seen as a highly dynamic medium and one that is much more
linked to the spread of changes than other written media and the archives we used• cover an extensive temporal range• include both large national papers and smaller local papers, thus
leading to a range of subjects, styles, and registers as well as regional variation.
Moreover, dictionary citations include hundreds of different CanE speakers, represented in smaller text fragments, giving greater (and more representative) usage
LMod
E-5
Typical BCE entry
LMod
E-5
Search screen
LMod
E-5
Three case studies• subjunctive in adverbial clauses –
a retention• progressive passive – an
innovation• modals and semi-modals – a
replacement
LMod
E-5
Subjunctive in adverbial clauses in PDEConditional clauses have provided “an extraordinary stronghold” for the subjunctive (Schlüter 2009: 281; also Harsh 1968: 42)Present subjunctive• very infrequent, limited to formal style, mostly realized as be• “obsolescent” in AmE (Algeo 2006: 38); “high-flown and probably
obsolescent” (Denison 1998: 294); a “SURVIVAL” (Fowler 1965: 596)Past subjunctive• distinguished from the past indicative only in the 1st and 3rd p. singular of BE
– the so-called were-subjunctive• used in hypothetical or unreal conditions, following if, as if, as though,
though, etc. (Quirk et al. 1985: 158), though the use with individual conjunctions varies widely (e.g. Visser 1972: 888ff.; Johansson and Norheim 1988: 33; Peters 1998: 97, 99; Grund and Walker 2006: 99; Schlüter 2009)
• associated with formal style (Quirk et al. 1985: 158, 1013, 1094; Peters 2004: 521; but cf. Leech et al. 2009: 66)
• fossilized in the if x were … construction (Fowler 1965: 595; Peters 1998: 101, but cf. Quirk et al. 1985: 1094; Leech et al. 2009: 65)
LMod
E-5
Subjunctive in adverbial clauses in PDEsigns of marginality, obsolescence – hypercorrections (use of the subjunctive in place of the indicative) – so-called “pseudo-subjunctives” (Ryan 1961; Fowler 1965: 597-598; Quirk et al. 1985: 158n.; Peters 1998: 96, 97; Algeo 2006: 39; Leech et al. 2009: 63, 63–64)• following if ‘whether’ in indirect questions: Danielle wondered if she were
getting enough to eat.• following if in open conditionals and temporal clauses: if this were his
intention, he failed to communicate to the control tower.declining use due to• redundancy (e.g., Peters 1998: 99; 2004: 521; also Visser 1972: 885;
González-Álvarez 2003: 306; Leech et al. 2009: 67)• formal syncretism of the subjunctive and the indicative (e.g., Leech et al.
2009: 67)continued use ascribed to• prescriptivism (Peters 1998: 98; Leech et al. 2009: 62, 68–69, especially in
AmE)• support from the mandative subjunctive in AmE (Leech et al. 2009:68)
LMod
E-5
Regional variationTwo early corpus studies:• The present subjunctive (mainly restricted to be) and in formal contexts is
“infrequent” in adverbial clauses; the were-subjunctive is more frequent, highest in informative prose and fiction. But “[t]he limited evidence does not suggest that there are any differences between British and American English” (Johansson and Norheim 1988: 34, 32, comparing LOB and Brown)
• the use of the counterfactual were subjunctive is much stronger in Brown and LOB than in ACE … the use of subjunctives in hypothetical conditional clauses is on the wane in Australia” (Peters 1998: 99)
Two more recent studies:• Compared to other contemporary national varieties, BrE and AmE form the
two endpoints of a dialectal continuum … these extraterritorial varieties [Indian, Australian, and New Zealand] English all use the subjunctive to a higher extent and/or at an earlier stage than BrE”. (Schlüter 2009: 283)
• “the were subjunctive is definitely losing ground in hypothetical adverbial clauses. From a more global perspective, AmE turns out to be the conservative variety in this ongoing change and BrE, for once, is more advanced” (Leech et al. 2009: 67).
LMod
E-5
Subjunctive in adverbial clauses in CanE
In CanE, the subjunctive in (as) if and (as) though clauses alternates with the indicative in hypothetical conditions, but• it is “still quite common, especially in formal usage” • it is a “stylistic choice” used to express, e.g., politeness or
cynicism• it should not be used when if means ‘whether’ nor when it
means ‘when’ (real or repeated possibilities)• it is required in if I were you in Standard CanE; “in formal
writing and speech, Canadians never use the phrase ‘If I was you’”.
(Oxford Guide to Canadian English Usage, Fee and McAlpine 2007: 261, 559-60)
LMod
E-5
BE-subjunctives in CanE (Strathy Corpus)
Table 1: Frequency of indicative and subjunctive forms of BE in (as) if, (as) though, and unless clauses with first and third-person singular pronoun subjects in the Strathy Corpus of Canadian English
Cf. subjunctives in the British National Corpus (BYU-BNC) in the present: the subjunctive is equally infrequentin the past: the subjunctive is the minority form but still well represented:
(as) if he|she|it were 36.9%(as) if I were 43.3%
So CanE resembles AmE in showing somewhat higher subjunctive rates than BrE
LMod
E-5
(Brief) history of the subjunctive in adverbial clauses“very little is known about [the subjunctive in] adverbial clauses of condition, concession and negative purpose” (Schlüter 2009: 277)In OE there was “a good deal of vacillation” between the indicative and subjunctive (Kilboon 1938: 261; Visser 1972: 882–885)In early ME usage is “rather confused”, but there is a “striking increase” in the use of the subjunctive in the later ME period (Kilboon 1938: 263–264).Auer (2006) shows a rapid decline to 1700, but González-Álvarez (2003: 307) finds stable usage during the 17th century (in the CEECS2) By 1700, the moods are distinguished only in the 3rd p. sg. of lexical verbs and BE (Beal 2004: 85–86), and the subjunctive declines (but not entirely consistently)
LMod
E-5
Decline of the subjunctivethe decline “has continued to this day, reversed sporadically only by the tendency to hypercorrection in 18c and later teachers and writers” (Strang 1970: 209)• (debatedly) the subjunctive is associated with the formal register
(Görlach 2001: 122; cf. González-Álvarez 2003: 309; Auer and González-Diaz 2005: 324–325; Grund and Walker 2006: 94–95; Auer and Tieken Boon von Ostade 2007: 7)
• it is a sociolinguistic marker of “polite” usage(Auer and Gonzalez-Diaz 2005: 321; Tieken Boon von Ostade 2009: 84)
• it is characteristic of the usage of women(González-Álvarez 2003: 310–311; Grund and Walker 2006: 97–98; Tieken Boon von Ostade 2009: 85) in the 19th century
• the subjunctive is more-or-less restricted to the verb BE(Strang 1970: 209; Grund and Walker 2006: 101, who find 92% BE)
• the indicative, not modals, take over for the lost subjunctive(González-Álvarez 2003: 306; Grund and Walker 2006: 93–94, 103; Leech et al. 2009: 65–66)
• the past were subjunctive becomes dominant over the present be subjunctive, and inanimate subjects replace animate ones (González-Álvarez 2003: 308; Grund and Walker 2006: 102)
LMod
E-5
18th and 19th centuries
18th c. • the subjunctive is “slightly more frequent” than in PDE (Görlach
2001: 122)• Auer (2006) finds a slight increase over the course of the century
(from 24.1% to 25.8%) which continues into the early 19th c. – perhaps a lag after the publication of 18th c. grammars
• there is a corresponding drop in the use of the indicative19th c.• a marked drop occurs in the second half of the century (Auer 2006;
Tieken 2009: 84, who dates the decline from 1870) • by the late 19th c., the subjunctive is seen as rapidly dying out (Bailey
1996: 217)• at this time, the use of the subjunctive has dropped to 22.7% in
letters (González-Álvarez 2003: 305), 22.1% in Helsinki Corpus and ARCHER (Auer 2006: 43)
LMod
E-5
The effects of prescriptivism• The survival is “partly supported by the acceptance of Latin-based
rules of correctness” (Görlach 2001: 122)• Visser (1972: 886–887) notes that 18th and 19th c. grammarians
condemned the “incorrect” use of was for were, but he finds the usage going back to ME
• Auer (2006: 45, 47) finds that the influence of grammarians is “not especially successful” and is “limited”, perhaps only effective in preventing the increasing “improper” use of was for were (also Auer and González-Diaz 2005: 323)
• Lowth only indirectly contributed to the use of the subjunctive (Auer and Tieken Boon von Ostade 2007: 15)
• But it is normative grammars (the “climate of the time”) and linguistic sensitivity of social climbers which contribute to the use of the subjunctive (Auer 2006: 48; Auer and Tieken Boon von Ostade 2007: 15; Tieken Boon von Ostade 2009:85)
LMod
E-5
BCE: Results by century
LMod
E-5
BCE: Past vs. present subjunctive
The last example of the present subjunctive in the BCE dates from 1953:If the jam starts, or any part of it, or if there be even an indication of its starting, he is drawn suddenly up by those stationed above.
There is some evidence for the shift from inanimate to animate subjectsanimate: 25.0% (1800s) > 28.6% (1900s) > 34.8% (2000s)
LMod
E-5
BCE: Subjunctive vs. indicative
Figure 4: Subjunctive and indicative forms of BE (as a percentage of the total) in the BCE
1750-1799 1800-1849 1850-1899 1900-1949 1950-19990%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
% indicative% subjunctive
LMod
E-5
Signs of obsolescenceThe indicative occurs in hypothetical conditionals in place of the subjunctive:• that if any representation is made to my prejudice, I expect I will be
allowed an opportunity of defending myself (CONTE-pC Let3)• I never breathe freely when a horse seems tired; I always feel as if I
was committing a crime riding it (CONTE-pC: Dia1)• the Tow Line which if it was to break would end in certain
Destruction to all (BCE 1821)• a Soldier … said, that if he was wanted, he was ready to give up the
names (BCE 834)• if a separation was desired by any in the country it was by the hon.
and learned gentleman’s party (BCE 1836)• one item is always a glass of wine if there is any (BCE 1844)• if he was put into jail, he could pay his hundred dollars to the King
George people (BCE 1860)• The expression “non-treaty Indian” means … even if such person is
only a temporary resident in Canada (BCE 1887)
LMod
E-5
Signs of obsolescenceThe subjunctive occurs in open conditionals, or cases where if means ‘when’:• The Indian name of it is woman’s tongue, for they say if one leaf be
set in motion all the rest begin, and then there is no such thing as stopping them (BCE 1791–1792).
• Woe betide his fresh shaven visage, if it be upreared above the hatchway! (BCE 1829).
• if the student were diligent and did nothing else he might fill a scribbler in a month (BCE 1897)
• If a man were very ill, and all remedies had failed to heal him the wabeno might place him beside the fire before the ceremony (BCE 1935).
The subjunctive occurs in indirect questions:• I would like to to [sic] know if there be any prospect of offering a
sale (CONTE-pC: Let3)• to ask if it were usual in Canada to do as the Whitby team was
doing (BCE 1963)
LMod
E-5
Conclusion: The subjunctive in if clauses in the history of CanEHistorical trends in the BCE correspond to those identified in other studies• the present subjunctive be was already in decline in the 1800s, was
rare in the 1900s and became obsolete in the mid-20th c.• the past subjunctive declined in the 1900s but continues to be used
about 15% of the time• evidence of the marginality of the subjunctive appears already in the
late 18th c., with use of the indicative for the subjunctive in hypothetical conditionals, as well as hypercorrect uses of the subjunctive in indirect questions and temporal clauses
• however, CanE shows a sharper rise in the first half of the 19th c., with rates approximating those of BrE only in the first half of the 20th c.
• the percentage use of the subjunctive in present-day CanE seems to be higher than that found in BrE
LMod
E-5
What is happening in PDE?Leech et al. (2009: 64) – comparing LOB/FLOB, Brown/Frown, found American English to be “lagging behind” BrE in the loss of the were-subjunctive • significant decline in BrE (from 63.3% to 51.9%)• non-significant slight increase in AmE (from 73.4% to 73.7%)
Is this • a retention or “extraterritorial conservatism”, i.e. the result of
“colonial lag” (Trudgill 2004), • an innovation, or • a revival?
• Leech et al. (2009:68) argue that “the relatively strong status that the were-subjunctive has in AmE might not simply have to be attributed to straightforward colonial lag but a more complicated pattern of post-colonial revival” – they speculate that AmE may have had a brief revival of the subjunctive from 1940 and 1960 and is now following the lead of the other dialects in moving towards the indicative
LMod
E-5
20th c. CanE subjunctive use
Figure 6: Frequency (per 1000 words) of BE subjunctives (1st and 3rd p. sg.) in CanE from the 20th-21st century (BCE)
1900-1919 1920-1939 1940-1959 1960-1979 1980-1999 2000-20130
0.01
0.02
0.03
0.04
0.05
0.06
LMod
E-5
Test Case 2: The progressive passive
LMod
E-5
The progressive passive“one of the few grammatical innovations” of LModE (Aarts et al. 2012: 870) There is agreement that the construction arose in the late 18th century in private correspondence and met virulent resistance (see Anderwald forthc.: §3):• . . . that of the House of Commons was being debated when the post went
out (1772 J. Harris Let. 8 Dec. in Early of Malmesbury Series Lett. First Earl of Malmesbury (1870) I.264; OED)
• A fellow … whose grinder is being torn out by the roots (1795 Southey in C. Southey Life I.249; OED) (a 1756 example, s.v. scrag, is rejected by Denison 1993: 439)
The modal and perfect forms (e.g. the issue may be being debated, the issue has been being debated) were integrated only in 20th century:• the first examples are “artificial” (Denison 1993: 480–430, 1998: 157)• the constructions are still very rare (Mair 2006: 90; Leech et al. 2009:
137) and may have “not yet reached the status of a generally recognized idiom” (Visser 1973: 2446).
Mod
E-5
(Brief) history of the progressive passiveThe progressive passive is the result of “systemic pressure” leading to a “much more symmetrical” auxiliary system
(Kranich 2010: 118–9, 242; Denison 1998: 151; cf. Visser 1973: 2426)
Its rise coincides with:• a general increase in the use of the passive (Arnaud 1998), especially with
non-agentive, non-human subjects (Hundt 2004a)• the (virtual) loss of the passival (e.g. the issues are debating)
“the rise of the progressive passive can … be situated squarely in the nineteenth century”; the turning point was the period 1850–1870
(Anderwald forthc.; Smitterberg 2005: 129)
It arose first in informal texts, but quickly spread to more formal texts(Visser 1973: 2426–7; Hundt 2004b: 109; Smitterberg 2005: 131)
In PDE, it is most common in informational (factually based, semi-formal) genres such as newspapers and least frequent in fiction
(Smith & Rayson 2007: 137–8; Hundt 2007: 297-8; Leech et al. 2009: 137, 142; Smith & Leech 2013: 86)
It is more common and rising in BrE but “lagging” in AmE(Hundt 2004b:110; Leech et al. 2009: 124, 136–7; Smith & Rayson 2007: 136; Hundt 2009:
17; but cf. Smith & Leech 2013: 85-6, who find no increase from 1931–2006)
LMod
E-5
The progressive passive in CanEStrathy Corpus: 1217 examples (normalized frequency 243.2/million)
British National Corpus (BYU-BNC): 16472 examples (normalized frequency of 171.11/million) (cf. FLOB frequency of 175/million, see Leech et al. 2009: 138, Figure 6.6)
Figure 8: Frequency and distribution of progressive passives by genre in the British National Corpus (accessed 10 July 2013)
Figure 7: Frequency and distribution of progressive passives by genre in the Strathy Corpus of Canadian English
Cf. COHA frequency (23 July 2013): 111.99/million
LMod
E-5
The progressive passive in World Englishes
Figure 9: Progressive passives in world Englishes (based in ICE-corpora)(from Hundt 2007: 295)
LMod
E-5
Earliest progressive passives in BCE
*no examples of be being, been beingTable 5: Earliest examples of the progressive passive in the Bank of Canadian English
LMod
E-5
Raw frequency of the progressive passive in BCE
Figure 10: Raw frequency of progressive passives over time in the Bank of Canadian English
1800-1819
1820-1839
1840-1859
1860-1879
1880-1899
1900-1919
1920-1939
1940-1959
1960-1979
1980-1999
2000-2013
0
20
40
60
80
100
120
140
LMod
E-5
Normalized frequency of the progressive passive in BCE
1800-1819
1820-1839
1840-1859
1860-1879
1880-1899
1900-1919
1920-1939
1940-1959
1960-1979
1980-1999
2000-2013
0
50
100
150
200
250
Figure 11: Normalized frequency of progressive passives (per million words) over time in the Bank of Canadian English
LMod
E-5
Progressive passive in COHA
Figure 12: Progressive passive in COHA (accessed 10 July 2013)
Earliest example: The negroes are being educated more rapidly, in large portions of the South, than are the people known as “poor whites.” (1820 COHA:MAG)
• We find the same general pattern until the 1940s, but then a decline in usage from the 1950s onward, with a steep fall beginning in the 1990s (to 84 per million words in the most recent period).
• Proscriptions against the passive generally in AmE may be responsible for this decline (cf. Smith and Rayson 2007: 150; Leech et al. 2009: 136)
• It could be speculated that the much higher usage in CanE shows that, perhaps because of differences in the educational system, Canada was not affected by US proscriptions.
LMod
E-5
Distribution of progressive passives
Figure 11: Distribution (by person and tense) of progressive passives in the BCE, ICE-CAN, and Strathy Corpus
• The progressive passive is most common in the present tense: 69.3% present, 30.7% past (cf. Hundt 2007: 298; Smith and Rayson 2007: 136; Leech et al. 2009: 124) – “colloquialization” (Mair 2006: 183ff.)?
• The perfect and modal forms are consistently very low frequency • Notably, the progressive passive in the first-person present tense is
extremely rare, both diachronically and synchronically.
am bein
g
is bein
g
are bein
g
was bein
g
were bein
g
be bein
g
been bein
g0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
45.0
BCEICE-CanadaStrathy
LMod
E-5
Summary: Progressive passive• BCE, despite a relatively low number of citations from the earlier
periods, provides quite old examples of this rare structure (often within 50 years of the oldest known examples)
• BCE shows a rise in frequency quite similar to that shown by COHA for AmE (without the decline from the 1950s)
• the distribution by person of the subject in CanE has remained relatively constant over time.
• The form is more common in the present tense.• The rarity of the form in 1st person singular has never been noted: this is
likely due to the fact that the progressive passive occurs overall quite infrequently with personal pronoun subjects (9.9% of the time); “am being” is the only form that occurs exclusively with a personal pronoun subject, the others occur with noun subjects as well.
• Despite the fact that BCE citations collected not in order to provide a representative example of CanE—or, more importantly, to record syntactic structures—the results are encouraging
LMod
E-5
Summary: Progressive passive
Pratt and Denison (2000; also Denison 1998: 153–5) argue that the progressive passive was spread by the “Lake School” literary group (c. 1795–1830)––a social network in Milroy’s sense––where a general if “unrespectable” form was consciously used by a group of young iconoclasts• BCE shows that this form was used (although sparingly) in printed
texts in Canada from 1830s onward• In the University of Virginia text collection, we find use of the
progressive passive quite early (1780s) in American English in official correspondence by George Washington and Thomas Jefferson
• in COHA, the form occurs 15 times in the 1830s in printed texts (6 in FIC, 7 in NF, 2 in MAG)
The spread of this form to post-colonial Englishes in such a short period of time, and its use in a variety of written genres, suggest that it could not have been spread by a small literary group in Britain
LMod
E-5
Test Case 3: Modals and semi-modals
(Dollinger forthc.)
LMod
E-5
Modals of obligation and necessity in CanE
(table adapted from Dollinger forthc.)
LMod
E-5
Replacement of must by have to
LMod
E-5
Rise of dynamic modal need to
LMod
E-5
Restriction of should to root contexts
LMod
E-5
Comparison of four deontic markers
LMod
E-5
ConclusionsThis paper has been an attempt to address two questions:1. How useful is the BCE for the study of historical CanE in the LModE period?2. What does it tell us about the nature of CanE (is it innovative, conservative)
and where CanE is going?The answer to the first question seems to be clearly “yes” (with certain limitations)• The BCE patterns in parallel with much larger corpora (such as COHA), even for fairly
infrequent grammatical phenomena• It gives us information about the history of a particular post-colonial variety of English
which we wouldn’t otherwise have access to• Its design (many small texts) is successful and should be considered an asset, not a
weakness (see Dollinger 2006, 2010)• But, like the use of the OED quotation database as an historical corpus, its validity as a
research tool might be questioned by strict corpus linguists.“we can expect three major benefits from the Bank of CanE: firstly, more reliable Canadian data … secondly, insights into the spread of changes in world English, from a North American – and Canadian – perspective, and thirdly a more reliable source for establishing regional difference within Canada”(Dollinger 2010: 108)
LMod
E-5
ConclusionsThe answer to the second question is more complexFor the BE subjunctive in adverbial clauses (which is generally seen as a retention)• The rates of subjunctive use in CanE, as in AmE, are higher than those in BrE• but what looks like “lag” (i.e. lag in the loss of subjunctives) may not be• there is a “lag” in the 19th c. – with higher rates overall and a steeper increase in
the first half of the century (perhaps showing the greater effects of prescriptivism, colonial insecurity, or lag)
• but CanE catches up to BrE (i.e. falls to the same rate) in the first half of 20th c.• and there is some evidence of a “revival” perhaps in both AmE and CanE in the
second half of 20th c.For the progressive passive (which is a “true” innovation)• CanE has higher rates than even BrE (but so do NZE and InE) – so certain
post-colonial varieties seem to be taking the lead in this innovation• The use of the progressive passive in AmE seems to have been slowed by
proscriptions against the passive generally (to which CanE was immune?), but this needs further investigation
• The spread of the progressive passive by a small British literary group is brought into question (by both AmE and CanE data)
LMod
E-5
ConclusionsFor the loss of modals and rise of semi-modals (both innovative)• For the modals, AmE and BrE “have been developing along broadly
parallel lines”, with AmE slightly in advance, but for the semi-modals, the picture is more complex; spoken data shows that AmE is “the main driving force of change” (Mair and Leech 2006: 327-8)
• Dollinger (forthc.) concludes that “CanE and AmE behave similarly” (for modals of obligation and necessity) in real time, with some differences: e.g., CanE is more progressive in replacing epistemic must with have to
• Collins (2013) shows that the rise of have to in AusE is intermediate between AmE and BrE, but the decline of must is greater than in both
• Thus, in respect to the modals and semi-modals, each variety more or less charts its own course, making it difficult to speak about conservative or innovative varieties
LMod
E-5
CodaSpeaking more generally, Hundt (2009: 14) says, “the dichotomy of ‘colonial lag’ and ‘colonial innovation’ – especially when it is applied to features of post-colonial English … – implies a far too simplistic view … a seemingly conservative feature may actually be a case of ‘colonial revival’”.
Grazie!