Piakandatu ami

Dr. Howard P. McKaughan edited by

Loren Billings and Nelleke Goudswaard

Exemplary analyses of the Philippine English Corpus Ma. Lourdes S. Bautista, De La Salle University-Manila (pp. 5–23)

Manila, 2007

ISBN 978-971-780-026-4

Exemplary analyses of the Philippine English Corpus Ma. Lourdes S. Bautista De La Salle University-Manila Exemplary analyses of the Philippine English Corpus

The inspiration for this paper comes from Schneider’s “Corpus linguistics in the Asian context: Exemplary analyses of the Kolhapur corpus of Indian English” (2000), illustrating how descriptive linguistics can be done using an electronic corpus. The general objective of his study was to characterize some features of Indian English and to determine whether this new variety was particularly conservative or relatively innovative compared to the colonizers’ variety, mainly British English, and peripherally American English. In other words, does Indian English manifest signs of a nativization process? His analysis of the Kolhapur corpus focuses on the use of the subjunctive, the case marking of whpronouns, the putatively British intransitive-do proform, and the indefinite pronouns ending in -body and -one. He compares his results with those of the (British) LancasterOslo-Bergen, hereafter LOB, and the (American) Brown corpora. The comparison seems well motivated because the three corpora—Kolhapur, LOB, and Brown—each consist of one million words, have an equal number of five hundred texts with two thousand words per text, follow the basic design of fifteen print genres or text types, with all the texts being synchronic: i.e., printed within the same year within each corpus (although it was 1961 for both Brown and LOB but 1978 for Kolhapur). I am replicating Schneider’s (2000) analysis of the subjunctive, the case marking of wh- pronouns, and the indefinite compound pronouns in -body and -one using two corpora—the components of the International Corpus of English (ICE) from the Philippines and Singapore, hereafter ICE-PHI and ICE-SIN. I have included the Singapore component * The availability in 2004 of the Philippine component of ICE has facilitated systematic studies of the lexical and grammatical features of Philippine English, in itself and in comparison with other corpora in ICE. ICE-PHI is an electronic lexical corpus of Philippine English consisting of three hundred spoken and two hundred written texts, with each text having approximately two thousand words, comprising a little over a million words. The Philippine corpus, together with corpora from six other countries or regions (Hong Kong, East Africa, Great Britain, India, New Zealand, Singapore), can be accessed from . With the use of a concordancing program such as WordSmith, electronic corpora present immense possibilities for interested students of Philippine English. It becomes easier, then, to implement a recommendation of Howard McKaughan, this volume’s honoree, from his article “Toward a standard Philippine English”: While the Philippines stands high amongst its peers in Southeast Asia in research on its own variety of English, there still remains much to be done. Teachers of language and linguistics should strive for even more advances in the study of Philippine English and to develop more efficient ways to teach Philippine English so that, beside the widening distribution and use of Filipino, the National Language, they will help Philippine English continue to flourish. (McKaughan 1993:53) I would also like to thank De La Salle University-Manila for the grant from the Research Faculty Program that enabled me to prepare this paper. Loren Billings & Nelleke Goudswaard (eds.), Piakandatu ami Dr. Howard P. McKaughan, 5–23. Manila: Linguistic Society of the Philippines and SIL Philippines, 2007.

6

MA. LOURDES S. BAUTISTA

because, like the Philippines, Singapore has a colonial history but, unlike the Philippines, it was under the British rather than the Americans.1 It should be pointed out that dates of compilation of the various corpora are different: The Kolhapur corpus was compiled in 1978, whereas ICE-PHI and ICE-SIN were compiled throughout the 1990s, with some data in ICE-PHI obtained as late as 2004. More critically, the Kolhapur corpus of one million words is drawn from printed material (500 texts times 2,000 words per text), while the ICE-PHI and ICE-SIN corpora of one million words each come from three hundred spoken texts (» 600,000 words) and two hundred written texts (» 400,000 words). Since the difference between spoken and written corpora has been found to be wide-ranging and considerable, in order to give greater validity to the comparison with Schneider’s findings, I have removed all the spoken text types together with the nonprofessional and correspondence genres. Thus, the investigation here has been limited to the published data (= 150 texts or 300,000 words), from ICE-PHI and ICE-SIN. To make the figures from Schneider’s paper comparable to the ones for ICE-PHI and ICE-SIN, I have used the normalization procedure described in Biber et al.’s Methodology Box 6, Norming Frequency Counts: normalization is “a way to adjust raw frequency counts from texts of different lengths so that they can be compared accurately” (1998:263–264). In this case, following Schneider’s advice (personal communication, April 2005), I have chosen one million words as the basis for norming, and this means dividing the raw frequency count for each feature by three, and then multiplying it by ten; the formula comes from the corpus size of 300,000 words normed to a million words ([300,000 ÷ 3] . 10 = 1,000,000).

1. Use of the subjunctive The SUBJUNCTIVE, as Schneider notes, is defined as a formally marked grammatical category: a distinct form of verbs with the meaning broadly being counterfactual. He writes that it is strongly recessive in the history of English; however, it has been retained in American more than in British English (2000:124). In British English, the subjunctive sounds formal and rather legalistic in style but it appears to be re-establishing itself perhaps because of American influence (Quirk et al. 1985:157). Thus, the subjunctive is an interesting structure for comparative studies because the two dominant first-language varieties (British and American) show different usage of the structure. As in Schneider, the analysis here focuses on the MANDATIVE subjunctive: the subjunctive used in that-clauses after verbs and, very occasionally, after SUASIVE expressions such as nouns and adjectives expressing a demand, order, recommendation, suggestion, or wish. 1

The bulk of the analysis for this paper was completed in mid 2005. Later that year (November), I was surprised to discover that Edgar Schneider’s contribution to a festschrift in my honor was entitled “The subjunctive in Philippine English”: I became anxious that his article had made my paper obsolete. However, though he covered basically the same ground that I have here, there are enough differences in the analysis to make publishing this paper worthwhile. For one, the interested reader can note the methodological differences between the two papers. Schneider used the entire ICE-PHI (both spoken and written subcomponents), whereas this paper has included only the printed dataset and used norming to approximate the one million words in the other printed corpora. For another, this paper has included data from ICE-SIN and therefore adds a dimension to the comparative analysis. Finally, the reader can observe slight differences in detail by comparing frequencies for the suasive verbs, adjectives, and nouns triggering the subjunctive and frequencies of the hypothetical subjunctive were with as if, as though, even if in the two papers. The important thing, however, is that we have both arrived at the same conclusion for the subjunctive and the equivalent should form—i.e., that Philippine English is highly predisposed to using the subjunctive rather than should and that it adheres very closely to American English in this regard.

EXEMPLARY ANALYSES OF THE PHILIPPINE ENGLISH CORPUS

7

The formal characteristics of the subjunctive have been succinctly described by Schneider (2000:124), summarizing Quirk et al. (1985:156[–157]): In many environments, a formal distinction between indicative (“base”) and subjunctive verb forms has been lost. The copula is the only verb which throughout its paradigm still has a distinct subjunctive form, viz. be. For all other verbs, the subjunctive equals the base form of a verb, i.e. it is distinctly recognizable (formally different from a non-subjunctive) only in certain environments: in the 3rd person singular in the present tense (because the subjunctive form lacks the verbal -s); in dependent clauses after past tense verbs in the main clause (because there is no tense concord, or “backshifting”); and, finally, in negatives (recognizable by the form not + infinitive without do-support). The most common functionally and semantically equivalent alternative of the subjunctive is the use of the modal should, said to be more common than the subjunctive especially in British English. The search structure for the subjunctives in ICE-PHI and ICE-SIN printed data followed that laid down by Schneider (2000:124–125) and reproduced almost verbatim below. As he notes, corpus linguistics is built on the need to identify a search structure so as not to have to read the whole corpus (an impossible undertaking); thus it is possible that this search structure missed a few tokens of the subjunctive, but they would be very few. To begin, define and formally identify the environments likely to trigger mandative subjunctives in dependent clauses—i.e., compile a list of suasive verbs and their derived nouns, such as demand, order, insist(ence), suggest(ion), and recommend(ation), as well as adjectives with a suasive meaning like imperative, important, and essential. The list used here is based on Schneider (2000) and Quirk et al. (1985:157, 1182). In addition, identify and clearly define the alternative formal realization categories (possible variants). Schneider identifies six of these, listed in (1) through (6)—corresponding to his types (a) through (f), respectively. Two of these—in (1) and (3)—are particularly important to Schneider’s study and the last two—in (5) and (6)—are marginal; he includes them only for the sake of completeness. Examples illustrating the six structures from the Philippine and Singapore corpora are listed within each category.2 (1)

a. b.

Distinct subjunctive: Those instances in which the verb of the complement clause is unambiguously subjunctive in form—i.e., the form be (e.g., We insist that he be informed), an uninflected third-person, singular form in a present-tense context (e.g., We insist that he go), or an uninflected verb of any grammatical person after a past-tense verb (e.g., We insisted that they go). It might be noted that the US is fairly transparent about the reasons for its insistence that agricultural liberalization be central to the APEC agenda (PHI, w2b–018) But they must plead as well that he show respect and devotion to that office. (PHI, w2e–001)

c.

2

We are not in any way asking that government restrict the movement of the foreign journalists (PHI, w2e–008)

The coding consists of the letters indicating the corpus (PHI = Philippines, SIN = Singapore) followed by the text type category (for example, w2b refers to written [printed] popular informational texts) and the text number (in the first case, 018 is from the social-science texts, since texts w2b–011 through w2b–020 come from that category). A full description of the compilation process of ICE-PHI and the different text types in ICE-PHI can be found in Bautista (2004). Italics have also been added to each quoted token.

8

MA. LOURDES S. BAUTISTA

d. e. f.

If there is no maintenance clause, the wife may later find it difficult to persuade the court to order that she be paid maintenance (SIN, w2b–002) The society proposed that four other sites be conserved instead: (SIN, w2c–005) She demanded that Uncle Ping take immediate action and stop Lienhwa at all costs. (SIN, w2f–011)

(2)

a. b. c.

Nondistinct, ambiguous forms: In present-tense contexts, unmarked verb forms in the first- and second-person singular and in all plural persons could be analyzed as either indicative or subjunctive forms (e.g., We insist that you go). Obviously, for a systematic analysis of the propensity of a variety to use subjunctives, these forms, being indeterminate—neutralized, as it were—cannot be used. he proposed that government and private business join hands in a conscious, coordinated effort (PHI, w2a–017) It is true that reason demands that we defend our laws and our institutions if we must remain a community (PHI, w2b–012) Senate probers proposed that instead of charging consumers a flat rate, PLDT and other companies charge subscribers on how often and how long they use the phone. (PHI, w2c–007)

d.

she insisted that all three of us get an [sic] university education and her reason is (SIN, w2b–013)

e. f.

Despite this, Dr. Kuek advises that patients have both operations done simultaneously. (SIN, w2c–015) Environmentalism as an idealistic preoccupation of a few makes it especially important that the eco-activists find domestic roots for a green Singapore. (SIN, w2e–002)

(3)

a. b. c. d. e. f.

(4)

a.

The modal should as the finite part of a dependent predicate: This is the dominant alternative strategy if the subjunctive is avoided, a grammatical pattern which is generally accepted and fully equivalent in meaning (e.g., We insist that he should go). The Department of Interior and Local Government has been unable to enforce its position that Cuneta should sit as Pasay City mayor. (PHI, w2c–018) He cited an executive order decreeing all contracts involving £ 50 million and above should undergo the approval of the Office of the President (PHI, w2c–020) Section 5 of RA 9048 and Rule 8 of Administrative Order No. 1, S. 2001 require that the petition should be in the form of an affidavit (PHI, w2d–010) Accordingly, we propose that the core value of “community over self” should be modified (SIN, w2a–012) and the other, by Virchow who insisted that disease should be localised in the body as lesion. (SIN, w2a–015) The NSS also suggested that once the nature areas are finalised, they should be legislated as such to ensure their protection. (SIN, w2c–005) Modals or quasimodals other than should (e.g., We insist that he must go): In corresponding structurally to should and conveying the same semantic component of counterfactuality, these are similar and might also be regarded as alternatives to the subjunctive; but any individual modal may add some meaning component of its own, so it is not necessarily clear that these are fully equivalent to a plain subjunctive. The authorization, they said, should clearly stipulate that the military authorities will allow access to all parts of the atoll (PHI, w2b–022)

EXEMPLARY ANALYSES OF THE PHILIPPINE ENGLISH CORPUS

b. c.

9

Lopez cites Kintanar for insisting to media that potassium bromate can be considered safe for the millions of people involved (PHI, w2b–028) Just quietly set an example and pray it will inspire them to do the same. (PHI, w2d–014)

d. e. f.

However, Wittgenstein’s insistence that the meaning of a word must be sought in its use points to a fundamental fact about language (SIN, w2a–005) Let us pray to God that such a calamity will never happen again and that He will prevail on our leaders (SIN, w2b–009) More important, all the republics are bound by a system that commanded that just one or two factories would supply the entire Soviet economy (SIN, w2e–004)

(5)

Negated dependent clauses with not + base form of the verb—i.e., without do-support (e.g., We insist that you not go): This is an extremely rare but structurally possible way of expressing a subjunctive. The only token comes from spoken data, in ICE-PHI. to preserve the just order and serve the common good we pray and pray intensely that we not be found wanting (PHI, s2b–023)

(6)

Clearly indicative forms (e.g., We insist that he goes), in which the modal (suasive) component is simply not expressed: These forms are also relatively infrequent. JAMP insists that there are other alternative uses for NAIA that can better benefit the people of Pasay (PHI, w2c–018) From a woman’s point of view, it is important that a separation deed entitles her to maintenance. (SIN, w2b–002) It is usually recommended that the Chief Executive Officer (CEO) of the company chairs a committee comprising the heads of departments (SIN, w2d–014)

a. b. c.

In the following discussion, I first present Schneider’s findings for the Kolhapur corpus in order to situate the findings for the Philippine and Singapore datasets. Table 1 below reproduces Schneider’s table for the Kolhapur corpus. Note that Schneider does not include the figures for the last two structures (negatives and indicatives), being marginal and relatively rare. As Schneider points out, it is clear that the subjunctive is relatively common in Indian English, though nowhere as frequent as its structural alternative should. In general terms, then, it can be said that the Kolhapur corpus follows the prediction that Indian English will approximate British more than American usage. As for the Philippine and Singapore printed data, using the search structure outlined above and using one million words as the basis for norming the 300,000 words of the printed material in each Asian corpus, tables 2 and 3 were obtained for ICE-PHI and ICE-SIN, respectively, again omitting the figures for the negatives and indicatives, as in Schneider, because they are marginal and rare. Clearly, the subjunctive appears more prominently in ICE-PHI than it does in the Kolhapur corpus; at forty-one percent, it is almost double the twenty-one percent in the latter. Likewise, the eleven percent usage of should as the structural alternative to the subjunctive in ICE-PHI is less than a quarter of the forty-nine percent usage of should in the Kolhapur corpus. In this case, the prediction that ICE-PHI would follow the American more than the British English pattern is corroborated. Moving next to the distribution of the subjunctive and should in the Singapore corpus, it is worth considering whether the city-state’s colonial history makes the figures for ICE-SIN

10

MA. LOURDES S. BAUTISTA

Table 1: Mandative subjunctives and main structural alternatives after suasive expressions in the Kolhapur corpus (Schneider 2000:127) Suasive expression

Subjunctive

advise

should

1

1

demand

6

9

desire

3

2 2

2

move

1 3

2

propose

1

3

recommend

6

22

request

1

1

require

2

1

stipulate

1

2

1

2

23

10

suggest

9

urge

1

1

wish

1

1

anxious

1

essential

1 1

1

2 2

1

demand

2

3

insistence

1

6

2

1

2

request (N)

1

suggestion

3

wish

1

Total

1

1

necessary

order (N)

2

1

order

important

Nondistinct

4

ask

insist

Other modals

35 (= 21.3%)

81 (= 49.4%)

2

1

33 (= 20.1%)

15 (= 9.2%)

EXEMPLARY ANALYSES OF THE PHILIPPINE ENGLISH CORPUS

11

Table 2: Mandative subjunctives and main structural alternatives after suasive expressions in the Philippine printed dataset (300,000 words and then normed to 1,000,000 words) should

Other modals

decide

2

2

decree

1

1

Suasive expression ask

demand

Subjunctive

Nondistinct

1

2

enforce

1 1

insist

1

order

1

plead

2

pray

3

1

1

1

propose

1

2

recommend

5

1

require

1

1

stipulate

1 1

suggest

1

fitting

1

important

1

2

1

insistent insistence Total Normed

3

1 2 19

5

10

12

63 (= 41.2%)

17 (= 11.1%)

33 (= 21.6%)

40 (= 26.1%)

adhere more closely to British English usage. The figures in table 3 can be compared with those given in tables 1 and 2 above for Kolhapur and ICE-PHI, respectively. The thirty-nine percent for the subjunctive in ICE-SIN is almost identical to the forty-one percent for ICE-PHI, and the twelve percent usage of should in ICE-SIN is likewise almost identical to the eleven percent usage in ICE-PHI. Those figures are decidedly different from Kolhapur’s figures of twenty-one percent for the subjunctive and forty-nine percent for should. Singapore English seems to follow the American rather than the British norm. After obtaining the frequency of the subjunctive and should in Kolhapur, Schneider poses what he considers the more interesting question: How does Indian English compare with other corpora in terms of this feature? He lists the frequencies of the subjunctive and should in Indian

12

MA. LOURDES S. BAUTISTA

Table 3: Mandative subjunctives and main structural alternatives after suasive expressions in the Singapore printed dataset (300,000 words and then normed to 1,000,000 words) Suasive expression

Subjunctive

should

Other modals

advise

1

command

1

demand

3

insist

1

order

2

1 1

pray

2

3

recommend

2

require

2

suggest

4

wish

1

2

1 1

3 1

1 2

1

insistence

1

proposal

1

requirement

1

suggestion

Normed

1

4

imperative

Total

1

1

propose

important

Nondistinct

5

1 20

6

11

15

67 (= 38.5%)

20 (= 11.5%)

37 (= 21.3%)

50 (= 28.7%)

English with those obtained by other researchers: Johansson & Norheim (1988) for British and American English, Peters (1998) for Australian English, Hundt (1998) for New Zealand English (as well as newer British and American corpora, from 1991), and subsequently obtains a table of proportions (2000:128). Table 4 is based on Schneider, to whose original table I have inserted the findings for Philippine English and Singapore English. Schneider makes several observations accompanying his original table (i.e., without the data from ICE-PHI and ICE-SIN in the bottom two rows of table 4). First, the expectation that the subjunctive is the normal choice in American English is borne out, as is the expectation that it would be rare in British English. However a comparison of the 1961 and 1991 British English data shows that the subjunctive gained in use, a fact noted by Quirk et al. (1985:157)—and quoted above in this paper—that the subjunctive seems to be re-establishing itself due to American influence. Second, Australian English shows a pattern close to that of

EXEMPLARY ANALYSES OF THE PHILIPPINE ENGLISH CORPUS

13

Table 4: Subjunctive vs. should in the different print corpora (Schneider 2000:128, with ICE-PHI and ICE-SIN appended) Corpus

Subjunctive

%

should

%

Indian English

Kolhapur

35

30.2

81

69.8

British English 1961

LOB

14

12.6

97

87.4

British English 1991

FLOB

44

39.6

37

60.4

American English 1961

Brown

116

85.9

19

14.1

American English 1991

Frown

94

89.5

11

10.5

Australian English

ACE

78

72.9

29

27.1

New Zealand English

WCNZE

70

66.7

35

33.3

Philippine English (normed)

ICE-PHI

63

78.8

17

21.2

Singapore English (normed)

ICE-SIN

67

77.0

20

23.0

Variety

American English, as does New Zealand English. Third, Indian English is considerably closer to British English than the other colonial varieties, with about thirty percent subjunctives compared to thirteen percent in British English and around seventy percent or more in the other varieties. The figures from Kolhapur made Schneider detect a certain tension in Indian English between the conflicting norms of British and American English (2000:129). Looking now at the new data for Philippine and Singapore English, it appears that both of these varieties follow American usage in the distribution of the subjunctive and should much more than the other colonial varieties, and are much further away from British usage. The Philippine results come closest to the American results, with the Singapore results just slightly different. The Kolhapur corpus was collected in 1978 and is considerably dated compared to the Australian, New Zealand, Singapore, and Philippine data. Thus, a study of the subjunctive in ICE-IND, released in 2002, should definitely be undertaken; constraints of time made it impossible to include that corpus in this study. However, Schneider quotes Hundt’s conclusion that American English “provides the model for World English ... [sic] in the revival of subjunctives after suasive verbs” (1998:165) and states that the Kolhapur data seemingly support this conclusion, showing that linguistic developments in India are in step with global trends (2000:129). It appears that the overarching influence of American English on norms for other Englishes is also the explanation for why the Singapore figures veer closer to the American figures than to the British ones. In addition to findings for the subjunctive, Schneider gives other evidence for the tension between competing norms in the Indian data. One piece of evidence concerns the negative-subjunctive construction cited in example (5) above, a structure classified as typically American (Schneider 2000:129, citing Johansson & Norheim 1988:30). He reports six occurrences in Brown, zero in LOB, and just one in Kolhapur—again suggesting a tendency towards, but not complete adherence to, British norms. In the new Asian ICE data, there was one occurrence of the negative subjunctive in ICE-PHI—the aforementioned (5)—and none in ICE-SIN.3 It is difficult, therefore, to draw any conclusions from the limited data. 3

Recall that (5) is from the spoken subcomponent, but note that this is the only time the whole corpus was searched (i.e., not just the printed data) because the printed data revealed no such occurrences.

14

MA. LOURDES S. BAUTISTA

Another piece of evidence was the use of the base subjunctive after the conjunction lest: zero in LOB, eight in Brown, one in Kolhapur (in addition to one token each of an indicative, a nondistinct form, and should). Thus, the Indian data show a skewing towards British English, but not complete conformity: a small step in the direction of American norms. In this regard, lest appeared once in ICE-PHI printed data, used with a subjunctive construction: (7)

Lest content be sacrificed, theological subjects are taught by professional theologians teaching in RSD. (PHI, w2a–003)

No such example was found in ICE-SIN printed data. Little can be said, then, about lest in ICE-PHI and ICE-SIN, except that the Philippine data showed one usage allied with American norms, not British. Schneider adds one more structure to his analysis of the patterning of Indian English with regard to the subjunctive: the use of was or were after the hypothetical conditional conjunctions as if, as though, and even if. Here, once again, Schneider devised a search structure that made it possible to survey the different datasets relatively quickly, by simply recording all instances of the occurrence of the conjunctions and seeing whether was or were had been used with them. Schneider’s results are presented in table 5, with the columns for ICE-PHI and ICE-SIN added. The two native varieties of Brown and LOB showed the highest percentages at eighty-one and sixty-five, respectively, followed very closely by the sixty-three of ICE-PHI. This may be an indication of Filipino speakers’ awareness of the need for the subjunctive in counterfactual statements—and this seems to be especially true for as though (with 77% were). In contrast, the other colonial corpora—Kolhapur at forty-nine percent, ACE at forty-six percent, and ICE-SIN at twenty-two percent—showed relatively fewer instances of the use of the subjunctive. In ICE-SIN, especially, was rather than were is the predominant form after the hypothetical-conditional conjunction. Table 5: Hypothetical subjunctive were vs. was in the different print corpora (Schneider 2000:130, with ICE-PHI and ICE-SIN added)

Conjunction

as if

was/were + was + were

Kolhapur

LOB

15 12

15 33

% subj.

as though

+ was + were

44 6 7

% subj.

even if

+ was + were

7 8

% subj.

Totals

+ was + were % subj.

28 27 49

65

17 15

1 19

3 10

4 4

13 57

77

50

0 13 3

0 10 17

46

26 3 0

0 0

30 26 81

20 7 50

44

43

ICE-SIN (normed)

7 7

9 7

4 3

ICE-PHI (normed)

47

95

41 34 62

ACE

81

71 10 7

53

8 35

69 9 22

54

Brown

19 36 10

63

22

EXEMPLARY ANALYSES OF THE PHILIPPINE ENGLISH CORPUS

15

In his study, Schneider offers a tentative hypothesis for the congruence in the findings for the colonial varieties Indian English and Australian English in the low use of the subjunctive with the as if/as though/even if conjunctions (and here we can perhaps now include Singapore English), as follows: “Colonial varieties tend to reduce grammatical complexity if it is not functionally required, i.e., meaningful” (2000:130). In short, for the counterfactual subjunctive, since no noticeable meaning is lost with the substitution of the indicative for the subjunctive, and since the subjunctive is grammatically quite complex, there might be a natural tendency to simplify. However, the Philippine data seem to refute that hypothesis, since the subjunctive appears more frequently than the indicative. Thus, it appears that no one explanation covers all the colonial varieties: For Indian, Singapore, and Australian English, it might well be simplification and the influence of British English; for Philippine English, it might be a tendency towards the more formal style, the style learned in the classroom (where, from observation, the subjunctive is thoroughly taught, or maybe even overtaught at the expense of other structures)—in addition to the influence of American English.

2. Case marking of wh- pronouns Schneider further analyzes the Kolhapur corpus in terms of the case marking of the whpronouns: i.e., the alternations between who and whom and between of which and whose (2000:130–131). Noting that English is moving away from being a largely synthetic language to a largely analytic language in its loss of case marking (as in, for example, the use of just uninflected who instead of the case-inflected distinction between who and whom in spoken English and the use of the analytic of which instead of the synthetic whose in the genitive case), he wanted to check manifestations of this in the Kolhapur corpus. His results are presented in table 6, together with the results that I obtained using the WordSmith concordancing program on the printed data of ICE-PHI and ICE-SIN. Table 6: Case marking of wh- pronouns in the different corpora (Schneider 2000:131, with ICE-PHI and ICE-SIN added) who

Corpus

whom

of which

whose

Kolhapur

6

(= 3.0%)

193 (= 97.0%)

75 (= 26.0%)

213 (= 74.0%)

LOB

8

(= 3.5%)

219 (= 96.5%)

122 (= 28.8%)

301 (= 71.2%)

Brown

2

(= 1.4%)

144 (= 98.6%)

78 (= 23.6%)

252 (= 76.4%)

17 (= 13.4%)

110 (= 86.6%)

133 (= 35.9%)

237 (= 64.1%)

90 (= 92.8%)

80 (= 31.1%)

177 (= 68.9%)

ICE-PHI (normed) ICE-SIN (normed)

7

(= 7.2%)

First, let us consider Schneider’s observations regarding the Kolhapur corpus in comparison with the L1-English corpora of LOB and Brown: The three varieties behaved very similarly, with the uninflected who occurring only around three percent of the time, and with the inflected whom prominently appearing in the three corpora. For the genitive, the synthetic whose appeared a little over seventy percent and the analytic of which less than thirty percent of the time. He again notes that the Kolhapur figures lie somewhere between Brown and LOB, slightly closer to British English with respect to the who/whom distribution, and pretty much half way between the two L1 varieties in the genitive. He also points out that in the Kolhapur corpus, in contrast to Brown and LOB, who occurs only in the more prototypical uses: i.e., the more informal genres of the printed data, and in interrogative

16

MA. LOURDES S. BAUTISTA

rather than relative clauses. Thus, he raises the question whether a higher distribution in more prototypical usages is a characteristic of L2 varieties. Turning now to figures of ICE-PHI and ICE-SIN usage, in terms of percentages, who taking the place of whom appears much more often in these two corpora than in the other three corpora, and conversely whom appears less in ICE-PHI and ICE-SIN compared to LOB, Kolhapur, and Brown. The figures in the newer corpora on which the comparison is made have been normed to one million words; here are the five occurrences in ICE-PHI and the two occurrences in ICE-SIN. (8) a. b. c. d.

Total sample of objective-case who (obtained from the un-normed 300,000 words) “Bye and thanks for visiting!” waved the appreciative lady who I correctly surmised to be the patient’s wife. (PHI, w2b–015) “We are a tough team to scout because you don’t know who you want to stop,” said Cone. (PHI, w2c–011) If you were at one time a member of the Babysitters Club like I was, you would know who I am talking about. (PHI, w2d–014) and much more sad, the remembered memory of who she used to ride with, back when she thought the clouds’ reflection on John Hancock was beautiful. (PHI, w2f–019)

e. f. g.

Pao, uh-wait! Pao: who do you love? (PHI, w2f–020) Signed by all 13 CFC members, the letter did not say who Mr. Chiam wanted to censure. (SIN, w2c–012) The difficulty lies in where to begin and who to begin with. It is easier to use yourself as the main subject (SIN, w2d–020)

A look at the PHI tokens reveals their occurrence in rather informal contexts, in featurized articles or fiction, and in quotations (thus, spoken); this conforms to Schneider’s observation that he found the usage of who in the more informal genres of his printed material. In this respect, the two Singapore tokens appear a little more formal. Thus, though the figures seem higher for uninflected who in ICE-PHI and ICE-SIN compared to LOB, Brown, and Kolhapur, there still seems to be awareness in the Philippine and Singapore data of the contexts where who and whom operate. Note that contrary to Schneider’s findings for Kolhapur, only one token out of the five in ICE-PHI is in an interrogative, in (8e), as opposed to a relative clause, in all of (8a–d). However, both tokens from ICE-SIN, in (8f–g), are in embedded interrogative structures. Regarding the genitive construction, the use of the synthetic whose is slightly lower (at 64% in ICE-PHI and 69% in ICE-SIN) than the figures ranging from seventy-one to seventy-six percent in LOB, Brown, and Kolhapur. Conversely, analytic of which seems somewhat higher in ICE-PHI and ICE-SIN, above thirty-one percent, compared to the numbers of twenty-three to twenty-nine percent in the three other corpora. It may be that the ICE-PHI and ICE-SIN writers are more wary of using whose for nonhuman antecedents, and this is in line with the following observation from Quirk et al. (1985:769): Whose is used with either personal or non-personal antecedents: The man whose wallet he stole The house whose rafters were burned There is a feeling, however, that whose is more appropriate to personal antecedents, presumably because of its morphological relationship to who and whom, and some speakers feel uneasy about its use with non-personal antecedents.

EXEMPLARY ANALYSES OF THE PHILIPPINE ENGLISH CORPUS

17

In fact, an investigation of the data shows that only twenty-six out of the total seventy-one tokens of whose in the Philippine dataset (37%) and only sixteen out of the fifty-three tokens in the Singapore dataset (30%) had nonhuman antecedents. Unfortunately, I have no easy access to the LOB, Brown, and Kolhapur corpora and therefore I do not have the percentage of nonhuman antecedents for whose in those corpora. It is certainly worth further study, in order to see whether the lower percentages for the synthetic whose and the higher percentages for the analytic of which in ICE-PHI and ICE-SIN are related in any way to animacy.4

3. Indefinite pronouns in -body and -one Again in replication of Schneider, I looked at the distribution of the pairs of indefinite pronouns everybody/everyone, anybody/anyone, somebody/someone, and nobody/no one. The forms in -one occur more frequently and are considered to express more elegance than the forms in -body; in addition, the forms in -body are more frequent in American than in British English. This section examines the distribution of these forms in ICE-PHI and ICE-SIN and how they compare with Schneider’s findings. Obviously, in each of the corpora in table 7, compound forms in -body are less frequent than those in -one. Schneider (2000:133) also cites Quirk et al. (1985:378), who make a similar statement accompanying their table showing frequencies of compound pronouns with any-, every-, and some- in British and American English. Quirk et al. add that, whereas the frequency of compounds in -body is not as high as that of -one compounds in American English, the frequency of -body compounds is higher in American than in British English, and that -one compounds are less frequent in American than in British English. Schneider echoes Quirk et al., noting that Brown has consistently more forms in -body than LOB does, and the Kolhapur data are in an intermediate position between the British and American English datasets but more closely aligned with British English. The new data from ICE-PHI and ICE-SIN—regrettably from much smaller datasets but normed to a million words—are remarkably similar to each other. The two of them show much lower percentages of compound forms in -body (the forms more commonly associated with American English) and conversely much higher percentages of forms in -one (the forms that “count as ‘more elegant’,” as Schneider 2000:133 writes, apparently not citing any other work). This could be a manifestation of an underlying awareness that compounds in -one sound more formal than those in -body and therefore are the preferred forms for printed texts.

4

An alternate analysis of whose and of which would probably require considering only the occurrences of nonhuman referents, for which both whose and of which would be appropriate. In itself, of which does not seem suitable as the pronoun for human antecedents. Consider the following example: [...] it was Mark’s opinion that John sought more. Mark whose labors in an advertising agency assured him of the wherewithal for his Sunday paintings (PHI, w2f–009). In this context, it is not possible to say, *[...] Mark, the labors of which in an advertising agency [...]. I consulted Edgar Schneider regarding this point and his reply (personal communication, April 2005) explains the logic for proceeding as he did in his study: [I]n modern English variability exists only with inanimate referents, so there are good reasons to restrict the comparison to such examples. I didn’t, however—basically because I had started out analyzing the paradigm in historical contexts, and there the restriction doesn’t apply; in fact, it’s one of the interesting topics to look into, as this semantic condition has grown over time since the Early Modern English period. [...] Schneider then proffers Our Father which art in heaven (from the “Lord’s prayer” in the King James Bible of 1611). To make comparisons with his results possible, I have therefore included all instances of whose and of which in this study.

18

MA. LOURDES S. BAUTISTA

Table 7: Indefinite pronouns in -body and -one in the different print corpora (Schneider 2000:133, with ICE-PHI and ICE-SIN added: both normed) Indefinite pronouns everybody everyone

Kolhapur

LOB

Brown

ICE-PHI

ICE-SIN

39 98

33 106

72 94

47 143

43 110

% -body anybody anyone

28.5 38 112

% -body somebody someone

32 141 25.3

23 107 % -body

nobody no one

60 118

Total -body Total -one

160 435 % -body

24.6

24.7 20 127

23.1 57 94

13 63

20 113

74 126

15.0

37.0

6.8 30 97

18.2 107 473

35.1

17.1 13 177

20 90

245 454

28.1

13.6

37.7

36.2 155 475

26.9

42 140

18.8 63 111

33.7

43.4

18.5 27 117

17.7

% -body

23.7

23.6 99 447

18.4

18.1

To test this hypothesis, a search was made of the -body and -one forms in the spoken data, consisting of 180 texts each from ICE-PHI and ICE-SIN (i.e., all of the more informal spoken texts, and therefore excluding the relatively formal ones of parliamentary debates, legal cross-examinations, business transactions, legal presentations, and speeches), each dataset totaling 360,000 words subsequently normed to one million words. These are shown in table 8. Indeed, it seems that compounds in -body are favored in spoken material. Certainly, much lower percentages of forms in -body were obtained in the printed data than in the spoken data (18% in printed vs. 49% in spoken for ICE-PHI; 18% vs. 61% for ICI-SIN, respectively), giving support to the statement that forms in -body are perhaps construed by users as more informal than forms in -one and therefore seem more suitable for the spoken mode. The pattern for the Philippine spoken data is more mixed than that for the Singapore data—preferences for everybody and nobody but for anyone and someone rather than their counterparts. In the Singapore spoken data, on the other hand, the preference is generally for the forms in -body, except in the compound someone. It appears, then, that the users of English represented in ICE-PHI and ICE-SIN have an implicit understanding that compounds in -body are more informal than compounds in -one, and thus are more suitable for the more conversational (spoken) genres. This conclusion is in line with the observation given above that forms in -one are somehow the “more elegant” form.

EXEMPLARY ANALYSES OF THE PHILIPPINE ENGLISH CORPUS

19

Table 8: Indefinite pronouns in -body and -one in selected ICE-PHI and ICE-SIN spoken texts (normed) Indefinite pronouns everybody everyone

ICE-PHI 222 122

% -body anybody anyone

64.5 53 103

% -body somebody someone

33.9 250 336

% -body nobody no one

42.7 94 75

% -body Total -body Total -one

55.6 619 636

% -body

49.3

ICE-SIN 197 61 76.3 81 50 61.8 181 217 45.5 131 44 74.8 590 372 61.3

4. Conclusion This study has analyzed data from ICE-PHI and ICE-SIN in light of Schneider’s findings concerning the Kolhapur corpus and its position in relation to LOB (the British corpus) and Brown (the American one). In order to make the results comparable across these different corpora, only the printed subcomponents of ICE-PHI and ICE-SIN were included in the comparisons. This meant, however, that the Philippine and Singapore data consisted of only 300,000 words, compared with the one million words of the LOB, Brown, and Kolhapur corpora. Norming was done to convert the ICE-PHI and ICE-SIN data to a million words, but the results might not be an accurate representation of the situation if a million words of print had actually been available in the PHI and SIN corpora. A number of results have been presented here, and the question can be asked: What are the patterns and trends that can be gleaned from these results? For Kolhapur, Schneider concludes that in all the features he studied—the use of the subjunctive, the case marking of wh- pronouns, the putatively British intransitive do, and the compound pronouns in -body and -one—the tendency was for the results to lie somewhere between those of LOB and those of Brown, with a slight tilt towards LOB. In short, Indian English does not share all the properties and distributions of its historical ancestor, British English; however, neither, apparently, does it deviate randomly, in an unpredictable and unordered fashion. Rather, he finds a pattern in the frequencies of the different features as leaning towards British English, though not identical with the frequencies of the parent variety. Schneider concludes in this way (2000:134): Clearly, it is possible to interpret these frequency distributions, which must have evolved over time like any other property of languages, as slow and gradual steps

20

MA. LOURDES S. BAUTISTA

towards linguistically indigenous preferences, i.e. to some extent the process of nativization, although this operates within a restricted framework of global varieties of English which are in consistent contact mutually. What, then, can be said of the findings for the two new Asian corpora? Based on the analysis presented here, the distributions for ICE-PHI and ICE-SIN, firstly, hang closely together, and, secondly, follow the American pattern in some respects but not in others. Let us now look closely at these claims. Concerning the subjunctive, ICE-PHI and ICE-SIN both use it preponderantly over should, and show figures approaching those of Brown and departing very much from those of LOB. In this general area of the subjunctive, going to the finer details, ICE-PHI adheres much more closely to Brown than ICE-SIN does in the use of subjunctive in the negative, with lest, and with as though. ICE-SIN shows some traces of its British origins in the use of the indicative rather than the subjunctive with the conjunctions as if, as though, and even if. The impression created by the results for the use of the subjunctive in ICE-PHI is one of rather strict adherence to prescriptive rules of American English. With regard to the use of who in the objective case, and the use of the synthetic whose and the analytic of which, ICE-PHI and ICE-SIN again seem to fall in together, both showing a tendency to depart from Brown and LOB, manifesting a relatively higher frequency distribution than the other corpora in the use of who in the objective case, and a lower frequency distribution for whose and a higher frequency for of which than the other corpora. A closer look at the examples of the use of who in place of whom in ICE-PHI shows that the contexts in which they appear are rather informal. On the other hand, a closer look at the use of whose in ICE-PHI and ICE-SIN seems to indicate a reluctance to use whose for nonhuman antecedents, perhaps explaining the lower frequency compared with the other corpora. However, since no similar close look could be done with the other corpora, it is not possible to make any further conclusions. In addition, because the corpus size is only 300,000 words, observations such as this should be taken with caution. As for the indefinite pronouns in -body and -one, compounds with -body show much lower figures for ICE-PHI and ICE-SIN compared with the other corpora; there was a more marked preference for -one in ICE-PHI and ICE-SIN compared to LOB, Kolhapur, and Brown. In this particular feature, ICE-PHI and ICE-SIN depart very much from Brown, showing only about half of the occurrences of -body and consequently manifesting much more use of the purportedly “more elegant” compounds with -one. A check of the ICE-PHI and ICE-SIN spoken texts, however, revealed much higher percentages of -body, especially for ICE-SIN, perhaps indicating some measure of awareness that compounds with -one are more formal in tone. Can a pattern be discerned in all these? Schneider (2000) has found a very neat pattern for his Kolhapur corpus: frequencies falling between the British and the American distributions, leaning slightly towards the British norms. The pattern is not so neat for ICE-PHI and ICE-SIN. There is a close adherence to American norms for the subjunctive (rather than should), on the one hand, but on the other, there is a lower frequency for whose and a higher frequency for of which, and a lower frequency for -body forms. I will not presume to provide an interpretation for the Singapore results and will leave it to Singaporean colleagues to take it from here.5 However, for the Philippine results, I tend to 5

The low figure for ICE-SIN, in contrast to that for the Philippine corpus—(22% vs. 63%)—certainly deserves further study. In discussion with this volume’s editors, an explanation was suggested to me that might be found in the L1s of the two populations: Tagalog and the other Philippine languages have very clear ways of expressing counterfactuality through verbal morphology, but the Chinese languages (mainly Mandarin and Southern Min) that are prevalent as L1s in Singapore do not.

EXEMPLARY ANALYSES OF THE PHILIPPINE ENGLISH CORPUS

21

see a pattern: Taken together, the features appear to cluster around the observance of prescriptive rules. The interpretation I would like to offer is that Filipino writers represented in this corpus seem to have acquired the rules for the use of the subjunctive quite well (and feel it better suited than should for print), have learned to associate whose with human antecedents and substitute of which for nonhuman antecedents, and have intuited that compounds with -one (rather than with -body) are “more elegant” and therefore are more appropriate for printed material. Perhaps these results are to be expected because the data come from the printed subcomponent of ICE-PHI, and therefore—of their very nature—are more formal. The only feature that does not follow this [+formal] tendency is the higher frequency of who in the objective case compared to the other corpora. But the contexts for the use of who in the Philippine data appear relatively informal, as exemplified in (8a–e) above, and therefore do not invalidate the claim of adherence to prescriptive rules. In fact, this observance of prescriptive rules is noted by Nelson (2005) in his study of the expression of future time, especially the use of shall, in ICE-PHI. He observes that shall in the Philippine data was reserved “almost exclusively for formal, public, non-interactive discourse, especially in ‘legalistic’ contexts” (2005:56) and that it was hardly ever used the way that British speakers and writers in ICE-GB used it, in informal, private contexts (conversations, phone calls, or social letters) as well as in “the ‘canonical’ formal context of administrative writing” (Nelson 2005:57).6 To explain this distinctive use of shall in ICE-PHI, which Nelson characterizes as “not ‘non-standard’ in any grammatical sense, but is certainly very distinctive from a stylistic point of view” (2005:57), he quotes Gonzalez (1991:334), who claims that formal style is the style that Filipinos are most comfortable with. Nelson (2005:57) adds: The fact that Philippine writers are more “comfortable” with the formal style would seem to imply that they are also more “comfortable” when they follow the prescriptive rules. This would seem to follow, too, from their alleged “stylistic insecurity”, since following the rules closely may be perceived as the best way to attain the appropriate style. Gonzalez attributes the stylistic underdifferentiation of Philippine English to the fact that Filipinos learn English in the classroom, and therefore formal-classroom English is the style they are most familiar with. Indeed, since the subcomponent of the ICE-PHI analyzed in this paper consists of published material, a majority of the writers included here almost certainly belong to the generation of English learners who acquired their English in the classroom and from personal reading of presumably quite formal material. The overly formal use of shall by Filipinos, as Nelson points out, may be “the result of ‘over-observance’ of prescriptive rules” (2005:58). Going further, this probably leads to the retention of certain conservative features of the colonizers’ language, an instance of “colonial lag” (Peters 1998:98, as cited by Schneider 2000:129). Observing the rules rigidly for certain grammatical structures, according to prescriptive grammar books, might be one route to the nativization of English in the Philippines, an institutionalization of the aforementioned colonial lag. Certainly, there is a need to study the spoken component of ICE-PHI to see whether the style indeed is still that of so-called classroom English or whether informality is more apparent in its grammar forms. The analysis of -body and -one compound pronouns in both the printed and spoken sections of ICE-PHI seems to indicate that differences exist and that the spoken subcomponent manifests more informality—i.e., higher use of -body forms. For one thing, 6

Personally, this comment struck me because Nelson’s description of the way Filipino users use shall is precisely the way I use it—always in formal contexts and never in private, informal, interactive contexts.

22

MA. LOURDES S. BAUTISTA

the same kind of comparison can be undertaken for the distribution of who and whom in the spoken and written subcomponents. The comprehensive grammar of Quirk et al. (1985) is a useful guide to grammatical forms that show distinctions in the formal and the informal registers. On the other hand, the kind of colonial lag alluded to above might well disappear in the not-so-distant future, given the high interconnectivity of the world’s population at this time. For the Filipino of today, the classroom is the whole world, and classroom English is the English not only of the Filipino teacher and the Filipino channel ABS-CBN but also of CNN, the Cartoon Network, and the Internet. Whether there will still be a tendency towards formality in Philippine English, as a result perhaps of the over-observance of prescriptive rules, remains to be seen. In summary, ICE-PHI is a ready resource for anyone interested in analyzing the Philippine English of the late 1990s and early part of the current decade and this study hopefully has indicated the possibilities for using it. In the interplay of varieties of English that mutually influence each other, the need will be to continually analyze and compare the different components of ICE and at the same time to update the corpus, including ICE-PHI, at regular intervals.7

References Bautista, Ma. Lourdes S. 2004. An overview of the Philippine component of the International Corpus of English (ICE-PHI). Asian Englishes 7(2). 8–26. Biber, Douglas, Susan Conrad, & Randi Reppen. 1998. Corpus linguistics: Investigating language structure and use. Cambridge: Cambridge University Press. Gonzalez, Andrew B. 1991. The Philippine variety of English and the problem of standardization. In Makhan L. Tickoo (ed.), Languages and standards: Issues, attitudes, and case studies. Singapore: SEAMEO Regional Language Centre. 86–96. Hundt, Marianne. 1998. It is important that this study (should) be based on the analysis of parallel corpora: On the use of the mandative subjunctive in four varieties of English. In Hans Lindqvist, Staffan Klintborg, Magnus Levin, & Maria Estling (eds.), The major varieties of English: Papers from MAVEN 97, Växjö 20–22 November 1997. Stockholm: Almqvist & Wiksell. 159–175. Johansson, Stig, & Else Helene Norheim. 1988. The subjunctive in British and American English. ICAME Journal 12. 27–36. McKaughan, Howard P. 1993. Toward a Standard Philippine English. Philippine Journal of Linguistics 24(2). 41–56. Nelson, Gerald. 2005. Expressing future time in Philippine English. In Danilo T. Dayag & J. Stephen Quakenbush (eds.), Linguistics and language education in the Philippines and beyond: A festschrift in honor of Ma. Lourdes S. Bautista. Manila: Linguistic Society of the Philippines. 41–59. Peters, Pam. 1998. The survival of the subjunctive: Evidence of its use in Australia and elsewhere. English World-Wide 19(1). 87–103. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, & Jan Svartvik. 1985. A comprehensive grammar of the English language. London: Longman. 7

In ways such as these, we will come closer to a Standard Philippine English, making some progress from McKaughan’s aim of moving toward such a standard language.

EXEMPLARY ANALYSES OF THE PHILIPPINE ENGLISH CORPUS

23

Schneider, Edgar W. 2000. Corpus linguistics in the Asian context: Exemplary analyses of the Kolhapur corpus of Asian English. In Ma. Lourdes S. Bautista, Teodoro A. Llamzon, & Bonifacio P. Sibayan (eds.), Parangal cang Brother Andrew: Festschrift for Andrew Gonzalez on his sixtieth birthday. Manila: Linguistic Society of the Philippines. 115–137. Schneider, Edgar W. 2005. The subjunctive in Philippine English. In Danilo T. Dayag & J. Stephen Quakenbush (eds.), Linguistics and language education in the Philippines and beyond: A festschrift in honor of Ma. Lourdes S. Bautista. Manila: Linguistic Society of the Philippines. 27–40.