E.
DOl: 10.4312/elope.10.1.9-21
Valentin Werner
University of Bamberg
he Present Perfect and Definite Temporal Adverbials: Reference Grammars and Corpus Evidence
Summary
he aim of the present paper is to assess the adequacy of how the grammatical status of the present perfect (PP) is established in three reference grammars of English (Quirk et al. 1985; Biber et al. 1999; Huddleston and Pullum 2002). I propose that the categorizations both as aspect and as tense as presented in these grammars have their inherent weaknesses and are particularly deficient when data from varieties other than British or American English is included. To test this assumption, I will analyze PP occurrences extracted from the International Corpus of English (ICE) appearing in contexts that have traditionally been considered ungrammatical or at least odd. One such context is the co-occurrence of the PP with definite temporal adverbials. In this case, the PP is used like a simple past, which may be taken as evidence for the tense status of the PP and will therefore be the focal point of the analysis.
Key words: present perfect, temporal adverbials, International Corpus of English, variation, tense, aspect, reference grammars
Dovršni sedanjik in določni časovni prislovi: zgledi iz priročnih slovnic in korpusa
Povzetek
Namen članka je oceniti ustreznost slovničnih opisov dovršnega sedanjika (Present Perfect) v treh priročnih slovnicah angleškega jezika (Quirk et al. 1985; Biber et al. 1999; Huddleston and Pullum 2002). Ugotavljam, da so klasifikacije dovršnega sedanjika, ki to strukturo uvrščajo tako v kategorijo slovničnega časa kot tudi v kategorijo glagolskega vida, neustrezne, še posebej takrat, kadar gre za nebritanske in neameriške variante angleškega jezika. Za potrditev te predpostavke bom analiziral pojavitve dovršnega sedanjika v International Corpus of English (ICE) s poudarkom na kontekstih, ki so v tradicionalnih opisih dovršnega sedanjika označene kot slovnično nepravilne oziroma nesprejemljive. Primer takšnega konteksta je raba dovršnega sedanjika skupaj s časovnimi prislovi, ki se nanašajo na določeno preteklo referenčno točko. V teh primerih se dovršni sedanjik rabi kot preteklik, kar lahko smatramo kot dokaz, da dovršni sedanjik sodi v kategorijo slovničnega časa.
Ključne besede: dovršni sedanjik, časovni prislovi, International Corpus of English (ICE), jezikovne variante, čas, glagolski vid, priročne slovnice.
he Present Perfect and Definite Temporal Adverbials: Reference Grammars and Corpus Evidence.
1. Introduction
Past time reference represents one of the notorious areas of English grammar, and the present perfect (in its standard form realized as have + V—en, henceforth PP) in particular seems to represent an elusive category. It has been given various labels such as 'aspect', 'tense' (Declerck 2006), 'phase' (Meyer 1992) or 'status' (Bauer 1970, cf. Kortmann 1995); the present study will focus on the first two categories as presented in influential reference grammars of English. To be precise, I will comment on the 'aspect' hypothesis, as posited by the Comprehensive Grammar of the English Language (CGEL; Quirk et al. 1985) and, in a similar fashion, the Longman Grammar of Spoken and Written English (LGSWE; Biber et al. 1999), as well as on the 'tense' hypothesis, as suggested in the Cambridge Grammar of the English Language (Huddleston and Pullum 2002).
A related point of interest, which has repeatedly been stated in the extant literature on the PP (e.g., Bauer 1970; Dahl 1985; Mugler 1988; Kortmann 1995; Portner 2003; Sempere-Martinez 2008; Ritz and Engel 2008, 138; Elsness 2009b or Schaden 2009), is that severe restrictions apply as to the combinability of the PP with temporal adverbials that refer to a definite point in the past,1 which render utterances like (1) ungrammatical or at least odd.
(1) *We've been at the meeting yesterday.
his constraint on combinability is known as the "present perfect puzzle" (Klein 1992; cf. Portner 2003) due to its typological oddity in contrast to perfects in other European languages. Note, however, that the restriction does not apply when non-finite forms, as in (2) or (3), or forms that are modified by a preceding modal, as in (4), are present (Giorgi and Pianesi 1997: 85).2
(2a) the process has now slowed down considerably, having passed its peak in the early 1970's (iCE-GB w1a-013)
(2b) Michael Heseltine having been through all that ten days ago went to his campaign office to work on maintaining and increasing his first ballot vote (ICE-GB s2b-003)
(3)	John is believed to have arrived at 2:00 yesterday (from McCawley 1971, 100; cited in Mugler 1988, 62-3)
(4)	his should've been put up a week ago!
1	hese include relative adverbials (e.g., yesterday, last week, two months ago; Schopf 1984, 16) that refer to a point wholly in the past.
2	Further contexts where the constraint is suspended do occur, e.g., with afterthoughts/insertions such as hey have also uhm I think last year uh given a list of of six principles (ICE-IRL s2b—001), in iterative contexts such as For example, the diagnostic and statistics manual (DSM) has been updated twice, once in 1968, and again in 1980, with a revised version appearing in 1986 (ICE-GB w1a—007; cf. Engel 1998, 132) or with postponed temporal specification due to pragmatic necessity as in Ive written a letter (speaker B) Have you (speaker A) Yeah last week (speaker B) (ICE-IRL s1a-059).
While some authors take it as self—evident that the relative frequency of the PP in contrast to the PT has constantly decreased in both British (BrE) and American (AmE) English (e.g., Elsness 1997; 2009a), others state that the variation between the two forms is diachronically stable in these varieties (Hundt and Smith 2009). Yet others claim that the PP is expanding its territory at the cost of the PT in some varieties, as it may be increasingly used in definite past time contexts (Engel 1998, 131). A number of corpus studies of BrE and AmE have identified instances of the PP with co—occurring definite temporal adverbials in both spoken and written material (Ota 1975, 46; Meyer 1995, 225; Hundt and Smith 2009, 55-7), which cannot all be simply attributed to performance errors or inadvertence on the part of the speaker (Quirk et al. 1985, sec. 4.23; Harder 1997, 417). It has been suggested elsewhere (Rastall 1999, 81-3; cf. Miller 2004, 235; Hundt and Smith 2009, 58 or Davydova 2011, 156-8) that the principle of current relevance might be superseding the combinational constraint explicated above for pragmatic reasons. As regards their overall frequency, examples of this kind are comparatively rare, but seem to be more salient and systematically used in other varieties, such as New Zealand English (NZE; Quinn 1999, 196; but cf. Hundt 1998, 74), Australian English (AusE; Ritz 2010, 3403; see further Ritz and Engel 2008) or various Asian Englishes (Kachru and Smith 2008, 91-2; Balasubramian 2009, 92-3).3 hus, it remains to be tested if and to what extent combinations of the PP and adverbials of the definite past group surface in the present dataset (see section 1 for details).
In the earlier literature there are numerous theoretical studies of the PP and the PP-PT contrast as well as empirical studies on this contrast and on temporal adverbials in what are conceived as the two traditional L1 varieties of English, BrE and AmE (e.g., Panzner 1995; Wynne 2000; Schlüter 2002). However, only few studies deal with the PP (e.g., van Rooy 2009) and the PP-PT contrast in further L1 and L2 varieties (e.g., Davydova 2011), and even fewer with the co-occurrence patterns of temporal adverbials in these other varieties (e.g., Hundt and Biewer 2007; Davydova 2011).
he present study can be seen as an attempt to complement earlier work. After a few comments on data and methodology, the first analysis section takes a closer look at the corpus data and tackles the following issues:
-	Is there evidence for the PP used as a "proper" tense that is on a structural/functional par with the PT in all or some varieties?
-	Are there qualitative differences between L1 and L2 varieties when combinations of PP and definite temporal adverbials occur?
his is followed by a contrastive view of grammatical categories as presented in three reference grammars of English in order to answer the following questions:
-	How is the PP embedded in models of grammatical categories in reference grammars?
-	What are the potential problems and weaknesses of these models?
In the concluding section, I will sum up the results and discuss further implications, such as the adequacy of current reference grammars in the area of the PP in view of the findings from the first section. I also address the need for potential modifications of future editions.
3 For the salience of the feature in other varieties see also Feature 100 (2012).
2. Data and methodology
he findings of the present study are based on data from the synchronic International Corpus of English (ICE), a family of comparable one million word corpora for a range of national or regional varieties of English. It contains educated English and includes both writing and speech.4 At present, ten components are publicly available5 and fourteen more are under compilation (http://ice—corpora.net/ice/index.htm). his corpus material allows for a meaningful comparative analysis of L1 and L2 varieties, which is only possible with parallel corpora that adhere to the same principles of design and thus are as homogeneous as possible with regard to text categories, date of production (early 1990s or later in the case of ICE) and the educational background of the speakers involved.
Relevant instances of PP + definite adverbial had to be retrieved from the corpus files, which had been part—of—speech tagged with the help of the C5 tagset of the CLAWS part—of—speech tagger (http://ucrel.lancs.ac.uk/claws/). With respect to the reliability of this tagger, the low error rate of 1.5% established for the tagging of the BNC was confirmed for the ICE data with the help of spot checks.6
he search was then undertaken with the help of the concordancing software WordSmith Tools (version 5.0.0.334; Scott 2008). It was for a selection of adverbials (x + ago, once, yesterday, last + x, in + cardinal number) that indicate a specific time in the past co-occurring with a PP (tags *_VHB/*_VHZ + *_VBN/*_VDN/*_VHN/*_VVN) and each individual corpus component was searched, leading to a file with concordance lines including a substantial amount of context (approx. 200 characters) before and after the individual occurrences for each component. Subsequently, the concordance lines were manually checked for unwanted occurrences (such as tagging errors or false positives), which were removed.
3. he present perfect and definite temporal adverbials
A corpus search for PP instances that violate the definite temporal adverbial constraint was conducted (cf. above); that is, occurrences that are considered ungrammatical or odd in standard BrE and AmE were identified in the ICE data. his allows us to test the hypotheses that (i) the PP is used as a proper tense that is on a structural/functional par with the PT and (ii) L2 varieties are more flexible in their usage of the PP in these contexts (see also Werner forthcoming). he search retrieved 57 relevant tokens in total. Figure 1 presents the counts for definite past adverbials co-occurring with the PP for the individual corpus components in the present dataset.
he first noteworthy finding is that instances of combinations of the PP and definite adverbials are attested across all varieties. Absolute numbers are low, which suggests that it constitutes a rare phenomenon in the majority of L1 and L2 varieties in contrast to observed frequencies of indefinite adverbials (cf. Werner forthcoming). Note also that other quantitative studies found that, overall,
4	he exact compositional principles of the corpus material are outlined in detail in a volume edited by Greenbaum (1996). See also http://ice-corpora.net/ice/design.htm.
5	hese are the components for Canada (CanE; ICE-CAN), East Africa (Kenya and Tanzania; EAfE; ICE-EA), Great Britain (BrE; ICE-GB), Honk Kong (HKE; ICE-HK), India (IndE; ICE-IND), Ireland (Northern Ireland and Republic of Ireland; IrE; ICE-IRL), Jamaica (JamE; ICE-JA), New Zealand (NZE; ICE-NZ), the Philippines (PhiE; ICE-PHI) and Singapore (SgE; ICE-SIN). Note that data from the Australian component (AusE; ICE-AUS), as kindly provided by Adam Smith (Macquarie University), and from a preliminary version of the Nigerian component (NigE; ICE-NIG; written genres only), as kindly provided by Ulrike Gut (University of Münster), were also included in the present analysis. Recently, the written part of the American component (ICE-USA) has also been made available.
6	See also van Rooy and Schäfer (2002), who found that the accuracy of the CLAWS tagger exceeded 96% for unedited and 98% for edited learner data where spelling errors were manually removed.
Figure 1. Co-occurrence of PP and definite temporal adverbials (absolute frequencies)7
less than one fourth of all PPs are specified by a temporal adverbial (e.g., Schlüter 2002) and the same applies to the ICE data (Werner forthcoming). Despite these low absolute numbers, some tentative conclusions can be drawn. First, PP co-occurrence is apparently more common in spoken data. his could be expected, as this mode has been found to be more susceptible to this pattern elsewhere (Hundt and Smith 2009, 55-6). Second, there is no significant difference between the proportions of spoken and written occurrences of the L1 and the L2 variety groups taken as wholes, as determined by a Fisher exact test. However, within the L2 group, the Asian varieties IndE and HKE and, to a lesser extent, PhiE are conspicuous as to their relatively high absolute frequency of PP co-occurrence with definite adverbials, while the remaining L2 varieties align with the L1 varieties.
here may be various explanations for these outliers. he PP has traditionally been prevalent in all contexts in IndE and it has been suggested that processes of transfer (reanalysis of the Hindi structure past participle + auxiliary and extension of its application to PP contexts) and substrate influence (for speakers of Dravidian languages) play a role for this variety (Davydova 2011, 172-3, 190), resulting in examples such as (5).
(5) He's been good yesterday (ICE-IND s1b-047)
In the absence of contradictory evidence and as no speaker information is available for the ICE components, this hypothesis seems to provide a plausible explanation. For HKE, it has been noted that its speakers commonly do not mark verbs for tense or aspect (or only draw a distinction between present and past tense), instead creating time reference through temporal adverbials alone (Platt 1982, 410; Setter et al. 2010, 49-55), as in (6) or (7), and that more variation in grammatical forms, including the PP with definite temporal adverbials as in (8) can be observed. In addition, forms that are partly marked (I've mention) as in (9) occur. All these items again are most likely due to a combination of substrate influence (Kachru and Nelson 2006, 42) and learner effects (Köppl 1983, 127) and surface also in the ICE-PHI data, as in (10)._
7 Although only written data are available for ICE-NIG, the example from this variety is categorized as spoken as it represents direct speech.
(6)	And then uhm but recently her backbone is very painful (ICE—HK s1a—052)
(7)	we just have a quarrel last night (ICE-HK s1a-035)
(8)	Some of them have once been my best friends (ICE—HK w1b—004)
(9)	So the most important thing when you look at the prologue okay is what as I've mention last time is the fact that it gives you a time structure (ICE—HK s1b—010)
(10)	Months ago I have written Sen John Sheffield the head of the US Senate Environment Committee about the problems (ICE—PHI s2b—032)
A further look at individual examples shows that, qualitatively, a diference between the occurrences in the L1 varieties and the L2 varieties can be established. While in both variety groups the subjective conception of a situation as recent and relevant by the individual speaker might be conveyed by a combination of definite and indefinite temporal adverbials as in (11) (see (12) for a corresponding PT example) or by a premodification (this last week, only yesterday) as in (13) or (14), the vast majority of occurrences in the L2 varieties clearly represents definite contexts that usually require the PT, as in (15) to (18). hus, the latter might well be explained by substrate and learner influence that results in establishing an innovative pattern, that is, the possibility of substituting the PT with a PP form in indefinite contexts, arguably leading to a conception of the PP as a proper tense that is on a functional par with the PT in these varieties.
(11)	I would not say that women's issues have just started last year I would not even say that (ICE— EA Kenya s1a—028)
(12)	I just had this kind of patient a couple of months ago (ICE—PHI s2a—032)
(13)	Oh I've had some fun this last week (ICE—CAN s1a—093)
(14)	Indeed my English is likely to be a little dodgy as I have only yesterday returned to civilisation after a 3—week sojourn to Derry (ICE—IRL w1b—006)
(15)	Actually I've noticed that two weeks ago (ICE—PHI s1a—066)
(16)	his work came out of a discussion that Noah and I have had several months ago probably in October uh last year when we were talking about ethics (ICE—JA s2b—029)
(17)	So he admits Sir what he has stated yesterday was was not correct (ICE—IND s2a—063)
(18)	Where have you been last Sunday (ICE—HK s1a—085)
For the L1 varieties, the few examples where PP co—occurs with definite temporal adverbials can be explained in terms of pragmatic necessity, as argued for instance by Rastall (1999: 81—3) and Hundt and Smith (2009: 58) (but cf. (19) and (20)). herefore, it seems unlikely that a general development towards PP—PT convergence/interchangeability will ensue in the L1 varieties in the foreseeable future.
(19)	Well he's come on very quickly last year (ICE—AUS s2b—017)
(20)	well because i mean we can read up on everything else again but we've seen that video months ago (ICE—NZ s1b—009)
4. he present perfect in reference grammars
As noted above, two reference grammars that have been very influential and are currently widely used are Quirk et al. (1985) and a more recent corpus-based work, Biber et al. (1999).8 herefore, a look at the statements on the PP contained in these major publications will be worthwhile. Both are congruent in that they recognize two tenses (present and past) and accordingly categorize the PP under aspect (Quirk et al. 1985, sec. 4.18; Biber et al. 1999, 452). Simultaneously, Quirk et al. concede that "aspect is so closely connected in meaning with tense, that the distinction in English grammar between tense and aspect is little more than a terminological convenience which helps us to separate in our minds two different kinds of realization: the morphological realization of tense and the syntactic realization of aspect" (Quirk et al. 1985, sec. 4.17). hus, the defining criterion for excluding the PP from the tense category is mainly formal in that only tenses can be structurally (i.e. morphologically) marked, an approach shared by Biber et al. (1999). Figure 2 illustrates the localization of the PP amongst the grammatical categories.
GRAMMATICAL CATEGORIES
tense
present
past
aspect
perfect(-ive) progressive -PP
other categories
voice, modality, (means of expressing future time)
Figure (G^amma^^ca^ cato^^o^ies ^^ r^resentt^d (Qhi^rk ^^ al.	(^ndBiber et al. (1999)
With regard to the distribution of forms with and without perfect aspect, Quirk et al. (1985, sec. 4.18) state that it can be found in around 10% of all VPs, a finding by and large reproduced by the data analysis by Biber et al., although there might be differences between individual registers (Biber et al. 1999, 461). In addition, both claim that perfects are more frequent in BrE compared to AmE (again with differences between different registers; Biber et al. 1999, 462-3; Quirk et al. 1985, secs. 1.24, 4.20).
Functionally, the two accounts take a slightly different approach. While both agree that the perfect aspect implies anteriorit:y (Qu^rk et al. 1985, sec. 4.18; Biber et al. 1999, 460), Quirk et al. (1985, sec. 4.20) describe the general meaning of the PP as expressing current relevance, while Biber et al. (1999, 460) emphasize the continuative and anterior properties of the construction. he latter refrain from further explanations of potential semantic readings, while the former can be seen as proponents of a polysemous account that lists three basic underlying meanings of the PP, namely (i) state leading up to the present, (ii) indefinite events in a period leading up to the present and (iii) habit or recurrent event in a period leading up to the present (Quirk et al. 1985, sec. 4.20). For the second reading three further implications are considered relevant: "(i) that the relevant time zone leads up to the present; (ii) that the event is recent and (iii) that the result of the action still obtains at the present time" (Quirk et al. 1985, sec. 4.21)._
8	For a discussion of the PP in older grammars see Vermant (1983, 12-5).
9	Means of expressing future are discussed in a separate chapter in Quirk et al. (1985, secs. 4.41-4.48), while they are included under modals by Biber et al. (1999, 483-97).
As regards the co-occurrence of the PP with definite temporal adverbials, Quirk et al. do not describe the phenomenon in detail. Surprisingly, they state that relevant instances surface "quite often" in BrE, but that these may mostly be seen as "performance errors" (1985, sec. 4.23). Although they recognize that other explanations might be needed, they do not comment on a potential dialectal influence.10
Further, both descriptions repeatedly stress the fact that the PP is interchangeable with the PT in some contexts, which strongly contributes to the "overlap of meaning between tense and aspect" (Quirk et al. 1985, sec. 4.18; cf. sec. 4.20) of the PP construction, while the diference in meaning between PP and PT may be attributed to temporal adverbials (Quirk et al. 1985, sec. 4.20; Biber et al. 1999, 467).11 Various factors are listed that might determine the speaker's choice between the two constructions, such as a more direct relation of the action to the present through the PP (Quirk et al. 1985, sec. 4.20), but the accounts remain at a superficial level.12
Huddleston and Pullum (2002) take a markedly diferent view of the PP hey also recognize "two primary tense categories, preterite (=past; V.W.) and present" (Huddleston and Pullum 2002, 125), but because they employ a wider definition of tense, which is inclusive of compound tenses (such as the PP), they class the PP as a "secondary past tense" (Huddleston and Pullum 2002, 116, 159; cf. also Huddleston 1996). hey are aware, however, that alternatives to their categorization such as 'perfect as aspect' or 'perfect as phase' exist. With regard to terminology, note that they use "past tense" as an umbrella term for both PT and PP forms (Huddleston and Pullum 2002, 116). his seems rather odd, in particular as they state later that the PP is indeed a combination of past and present and even that "in the present perfect the primary tense is present" (Huddleston and Pullum 2002, 142; see also Biber et al. 1999, 460), which is inconsistent with their general approach to grammatical categories (see Figure 3).13 hey further explicate their "past tense" umbrella category by stating that anteriority is the common denominator of both forms (139). heir overall model stands in stark contrast to the views expressed by Quirk et al. (1985, sec. 4.18) and Biber et al. (1999, 460). Specifically, for Huddleston and Pullum the defining feature of the PP compared to the PT is that the former is non-deictic (2002, 140-3), while the latter is deictic and "doubly anterior" as it "locates the writing anterior to an intermediate time which is anterior to the time of speaking" (140). he PP is only anterior to the time of speaking (142).14 herefore, in a rather imprecise manner, they see the PT as "a clearer instance of a tense" than the PP (159; emphasis added). he explanation of the choice criteria between the two constructions is again based on pragmatic notions such as speaker focus on either past or present (143).
10	A leveling of the difference between PP and PT, which probably is a remnant from older stages of English, has been attested for a number of regions of the British isles, such as East Anglia, Wales and Northern areas (Feature 100 2012).
11	In contrast to the considerations presented by Biber et al. (1999, 467-8), temporal adverbials with the PP are comparatively rare, i.e. less than one fourth of all PPs are specified (cf. above). See Declerck (2006, 316-26) for an attempt to devise a more elaborate description of possible selection criteria in sentences without temporal adverbials. He writes that "when, in a noninterrogative clause referring to a bygone situation, there is no time-specifying adverbial, both the preterite (= PT; V.W.) and the indefinite present perfect are in principle possible" and that pragmatic "speaker focus on Now or then" (2006, 322; emphasis original) is the crucial underlying notion between using either PP or PT (see also sec. 3 below).
12	Note that Biber et al. provide interpretations of their corpus data with regard to the most common verbs that occur in combination with the PP (1999, 464) and the special status of have got/gotten and have had as highly frequent possessive constructions (1999, 466).
13	Cf. also Huddleston and Pullum (2002, 143): "he possibility of having present time adjuncts like now or at present shows clearly that we have present time meaning as well as present tense form." However, the counterargument that temporal adverbials that (also) create reference to the past such as since X or previously may occur and that the PP therefore must (also) have a past time meaning can be considered equally valid. In addition, it seems redundant to propose a binary distinction (primary and secondary tenses) only to abandon it on a lower level of analysis and to substitute it with an ad-hoc category (past tense(s)) that is not anchored in the general model. Simultaneously, this once again highlights the rather elusive character of the PP.
14	Mutatis mutandis (with a change of perspective from anteriority to posteriority), this view is reminiscent of Mugler's approach, which states that the PP expresses simple posteriority while the PT expresses complex posteriority (1988, 235).
VERBAL SYSTEMS
tense	aspect	mood
primary secondary progressive non-	modal non-
past tens^^\	progressive	modal
present
past perfect
non-perfect
' PP
Figure 3. Grammatical categories (=verbalsystems) as represented by Huddleston andPullum (2002)
Functionally, a polysemous account is provided, surprisingly enough in the chapter focusing on the differences between PP and PT usage (143—6). It recognizes a twofold view with continuative uses, which are claimed to be usually accompanied by temporal adverbials (141—2), and non— continuative uses, namely (i) experiential/existential, (ii) resultative and (iii) recent past (see 143—6 for example sentences and elaborations on when these readings are possible). Note that Huddleston and Pullum concede that these classifications are not necessarily mutually exclusive (2002, 143). Similar to Quirk et al. (1985), a general current relevance notion is asserted for the PP, although no further details are given besides the fact that the PP is somehow concerned with the "time span up to now" (Huddleston and Pullum 2002, 143).
As to the co-occurrence of PP and adverbials with reference to a definite point in the past ("past time adjuncts" in their terminology), Huddleston and Pullum (2002, 143) state only that the PP (as a tense which refers to both present and past) is incompatible with adverbials that "refer to times wholly before now". Under certain conditions, however, past time adverbials may co-occur, as long as the reference to the point of speech ("now") is the focus of the utterance, for instance when another non-past adverbial, as in (21), is added.
(21) We've already discussed it yesterday. (Huddleston and Pullum 2002, 144)
What is clear from the contrastive view and particularly the visualizations is that the PP does not fit into the general models of grammatical categories without some kind of workaround. herefore, both of^ the pr^ncip^l appro ac;he^ (PP as aspect vs. PP as tense) represented in major reference grammars possess some inherent weaknesses: the former as it is solely form-based (see also Nordlander 1997: 16-8); the latter as it bears some internal inconsistencies, in particular by introducing the "past tense" label that comprises both past and perfect. A further issue which is detrimental to the internal consistency of Huddleston and Pullum's model is the finding that they describe "non-perfect" as constituent of their secondary tense system (2002, 116; see also Figure 3). Elsewhere, however, they state that "non-perfect is not a tense" (2002, 159). More generally speaking, both approaches as outlined in the preceding sections remain vague at times in their descriptions of the PP.
5. Discussion and conclusion
In the present paper, I have examined a number of PP examples from various varieties of English that violate the definite temporal adverbial constraint, and reported on the models of grammatical
categories in three influential reference grammars. Below, I will provide a brief summary and relate the findings of both sections.
As regards the weakening of the definite temporal adverbial constraint, it transpired that PP usage in definite temporal contexts occurs in the ICE data, albeit rarely overall15 and predominantly in spoken genres. A qualitative analysis revealed that while pragmatic pressure on the speaker as proposed by Rastall (1999) may override the constraint in L1 varieties, L2 speakers are more likely to show an innovative use of the PP - that is, to use it interchangeably with the PT - due to substrate or learner eflects or a combination of both. As co-occurrence of the PP and definite temporal adverbials is salient in older stages of English (Miller 2004, 235), it is also conceivable that it represents a residue that was introduced in early stages of colonization and has been preserved in postcolonial varieties (cf. Hickey 2004 on the "founder principle"). herefore, based on the present evidence it is conceivable that change in this grammatical area (development of the PP into a proper tense on a structural par with the PT, i.e. a fully-fledged variant) might occur (cf. the typological evidence in many other languages; Miller 2000, 350). Yet, while PP usage in combination with definite temporal adverbials may become sanctioned in one variety (e.g., one of the L2 varieties or in one of the "transplanted" L1 varieties such as AusE; see Ritz 2010), it remains speculative whether this development will gain currency over time and eventually spread to other varieties. It has to be conceded that at present the case for an extended functional range of the PP in the varieties under investigation is weak and no definite answer can be given as to whether this will materialize. Further work, for example involving longitudinal quantitative study of both L1 and L2 varieties, is needed to establish whether PP usage with definite temporal adverbials has been diachronically spreading or is rather less regular and might therefore be attributed to learner and substrate eflects alone.
With the corpus analysis as foil, we can now assess the adequacy of the models of categorization of the PP in the reference grammars. Broadly speaking, it emerged that two types of categorization (PP as aspect vs. PP as tense) are presented. Both models fail, however, to provide conclusive explanations as to why the PP may disregard the definite temporal adverbial constraint within their models. On the one hand, Quirk et al. (1985) ascribe violations merely to performance errors, an approach that does not seem to be universally applicable, particularly for L2, but also for some of the L1 varieties. hus, the restricted variational focus (Standard BrE and AmE only) that has been most explicitly criticized by Phillipson (2009, 31) is indeed detrimental to the value of the book as a "comprehensive" grammar. Besides, no statement on the issue of combinability or the occurrence of instances that violate the constraint can be found in the relevant section of Biber et al. (see 1999: 467-8), where findings are also based on BrE and AmE data only. On the other hand, the account by Huddleston and Pullum (2002) prima facie seems more adapted to the empirical findings, given that in their work the PP is analyzed as tense. However, it is flawed on logical grounds as the PP is categorized both as past and present tense. Furthermore, to label the PP a secondary (past) tense is somewhat infelicitous in the face of the fact that the PP in some varieties fulfills the same function as the PT (which is a primary past/secondary non-perfect tense), as has been shown above. Again, the situation as to the relaxation of constraints on the combinability of the PP in varieties other than BrE and AmE is disregarded.
In conclusion, I suggest that future editions of reference grammars that claim validity for English as a whole (like those discussed in the present paper, with exception of Biber et al. 1999) would do well to build on resources such as Kortmann & Lunkenheimer (2011) and to contain some (more) explicit
15 Against the objection that the low number of occurrences might be merely ascribed to the compilation principles of ICE, which captures mainly non-vernacular English, note that Davydova's (2011) data on mesolectal and basilectal varieties, which were obtained by way of sociolinguistic interviews, also did not yield sufficient numbers to attempt a meaningful quantitative analysis.
reference to the fact that the definite temporal adverbial constraint might be relaxed in some varieties and under certain pragmatic conditions, allowing tense-like usage of the PP. Admittedly, it might be argued that this feature is only restricted to the vernacular and therefore insignificant in general descriptive accounts of English grammar. However, it appears to be pervasive in a number of South and Southeast Asian varieties (Feature 100 2012) and occurs in the (predominantly spoken) language of educated speakers as contained in ICE, including L2s such as IndE or HKE as well as L1s such as NZE or AusE (see above). With hints to these findings grammars would, at the same time, take account of the primacy of speech, one of the long-standing axioms of linguistic study. Qualitative differences between the variety types would still have to be pointed out, of course. Further, it is already recognized that PP as a less and PT as a more grammaticalized variant in general "have a great deal in common" (Huddleston and Pullum, 2002: 116); thus, the (quasi-) interchangeability of the two forms in some contexts does not come as a great surprise. A large-scale corpus project focusing on grammatical features as mentioned above would be desirable in order to obtain definitive results. Seen from a more general perspective, while establishing a comprehensive unified grammar of "World English(es)" certainly constitutes a controversial project, extending the focus beyond the traditional BrE/AmE paradigm would do more empirical justice to English as a pluricentric language.
References
Balasubramian, C. 2009. Register Variation in Indian English. Amsterdam: Benjamins. Bauer, G. 1970. he English "Perfect" Reconsidered. Journal of Linguistics 6: 189-98.
Biber, D., S. Conrad and G. N. Leech. 1999. Longman Grammar of Spoken and Written English. Harlow: Longman. Dahl, Ö. 1985. Tense and Aspect Systems. Oxford: Blackwell.
Davydova, J. 2011. he Present Perfect in Non-native Varieties of English. Berlin: Mouton de Gruyter.
Declerck, R. 2006. he Grammar of the English Verb Phrase: Vol I. he Grammar of the English Tense System: A ensive Analysis. Berlin: Mouton de Gruyter.
Elsness, J. 1997. he Perfect and the Preterite in Contemporary and Earlier English. Berlin: Mouton de Gruyter.
---. 2009a. he Perfect and the Preterite in Australian and New Zealand English. In Comparative Studies in
Australian and New Zealand English: Grammar and beyond, eds. P. Peters, P. Collins and A. Smith, 89-114. Amsterdam: Benjamins.
---. 2009b. he Present Perfect and the Preterite. In One language, two grammars? Differences between British and
American English, eds. G. Rohdenburg and J. Schlüter, 228-45. Cambridge: Cambridge University Press.
Engel, D. M. 1998. A Perfect Piece? he Present Perfect and Passe Compose in Journalistic Texts. Belgian Journal ofLinguistics 12: 129-47.
Feature 100. 2012. Levelling of the difference between present perfect and simple past: present perfect for StE simple past. In he electronic World Atlas of Varieties of English, eds. B. Kortmann and K. Lunkenheimer. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available online at http://www.ewave-atlas. org/ parameters/100.
Giorgi, A. and F. Pianesi. 1997. Tense and Aspect: From Semantics to Morphosyntax. Oxford: Oxford University Press.
Greenbaum, S., ed. 1996. Comparing English Worldwide. Oxford: Clarendon Press.
Harder, P. 1997. Functional Semantics: A heory of Meaning, Structure and Tense in English. Berlin: Mouton de Gruyter.
Hickey, R. 2004. Introduction. In Legacies of Colonial English: Studies in Transported Dialects, ed. R. Hickey, 1—30. Cambridge: Cambridge University Press.
Huddleston, R. 1996. he English Perfect as a Secondary Past Tense. In he Verb in Contemporary English: heory and Description, eds. B. Aarts and C. F. Meyer, 102—22. Cambridge: Cambridge University Press.
---and G. Pullum. 2002. he Cambridge Grammar of the English Language. Cambridge: Cambridge
University Press.
Hundt, M. 1998. New Zealand English Grammar, Fact or Fiction? Amsterdam: Benjamins.
---and C. Biewer. 2007. he Dynamics of Inner and Outer Circle Varieties in the South Pacific and Asia.
In Corpus Linguistics and the Web, eds. M. Hundt, N. Nesselhauf and C. Biewer, 249—69. Amsterdam: Rodopi.
---and N. Smith. 2009. he Present Perfect in British and American English: Has here Been a Change,
Recently? ICAME Journal 33: 45—63.
Kachru, Y. and C. L. Nelson. 2006. World Englishes in Asian Contexts. Hong Kong: Hong Kong University Press.
---and L. E. Smith. 2008. Cultures, Contexts, and World Englishes. New York: Routledge.
Klein, W. 1992. he Present Perfect Puzzle. Language 68(3): 525—52.
Köppl, S. 1983. Englisch in Süd— und Südostasien. In Englisch, Formen und Funktionen einer Weltsprache, eds. S. Köppl, J. Schmied, E. W. Schneider and W. Viereck, 117—43. Bamberg: Lehrstuhl für Englische Sprachwissenschaft und Mediävistik.
Kortmann, B. 1995. Compositionality and the Perfect. In Anglistentag 1994 Graz: Proceedings, ed. W. Riehle, 183—99. Tübingen: Niemeyer.
Kortmann, B. and K. Lunkenheimer, eds. 2011. he electronic World Atlas of Varieties of English. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available online at http://www.ewave—atlas.org.
Meyer, M. 1992. Das englische Perfekt. Tübingen: Niemeyer.
---. 1995. Past Tense and Present Perfect in the Brown and the LOB Corpora. In Anglistentag 1994 Graz:
Proceedings, ed. W. Riehle, 201—28. Tübingen: Niemeyer.
Miller, J. 2000. he Perfect in Spoken and Written English. Transactions of the Philological Society 98(2): 323—52.
---. 2004. Perfect and Resultative Constructions in Spoken and Non—standard English. In Up and down the
Cline: he Nature of Grammaticalization, eds. O. Fischer, M. Norde and H. Perridon, 229—46. Amsterdam: Benjamins.
Mugler, A. 1988. Tempus und Aspekt als Zeitbeziehungen. München: W. Fink.
Nordlander, J. 1997. Towards a Semantics of Linguistic Time: Exploring Some Basic Time Concepts with Special Reference to English and Krio. Uppsala: Swedish Science Press.
Ota, A. 1975. Tense and Aspect of Present—day American—English. Tokyo: Kenkyusha.
Panzner, A. 1995. Das Pre—present in Kollokation mit Temporalangaben im britischen und amerikanischen Englisch der Gegenwart. Frankfurt am Main: Lang.
Phillipson, R. 2009. Linguistic Imperialism Continued. London: Routledge.
Platt, J. T. 1982. English in Singapore, Malaysia, and Hong Kong. In English as a World Language, ed. R. W. Bailey, 384—414. Ann Arbor: University of Michigan Press.
Portner, P. 2003. he (Temporal) Semantics and (Modal) Pragmatics of the Perfect. Linguistics and Philosophy 26: 459-510.
Quinn, H. 1999. Variation in New Zealand English Syntax and Morphology. In New Zealand English, eds. A. Bell and K. Kuiper, 173-97. Amsterdam: Benjamins.
Quirk, R., S. Greenbaum, G. N. Leech and J. Svartvik. 1985. A Comprehensive Grammar of the English Language. Harlow: Longman.
Rastall, P. 1999. Observations on the Present Perfect in English. World Englishes 18(1): 79-83.
Ritz, M. A. 2010. he Perfect Crime? Illicit Uses of the Present Perfect in Australian Police Media Releases. Journal of Pragmatics 42: 3400-17.
---and D. M. Engel. 2008. "Vivid Narrative Use" and the Meaning of the Present Perfect in Spoken Australian
English. Linguistics 46(1): 131-60.
Schaden, G. 2009. Present Perfects Compete. Linguistics and Philosophy 32: 115-41.
Schlüter, N. 2002. Present Perfect. Tübingen: Narr.
Schopf, A. 1984. Das Verzeitungssystem des Englischen und seine Textfunktion. Tübingen: Niemeyer. Scott, M. 2008. WordSmith Tools Version 5. Liverpool: Lexical Analysis Software.
Sempere-Martinez, J. A. 2008. Cultural and Psychological Factors for the Present Perfect in British English and American English. he International Journal of Arts Education 6(1): 123-39.
Setter, J., C. S. P. Wong and B. H.-S. Chan. 2010. Hong Kong English. Edinburgh: Edinburgh University Press.
van Rooy, B. 2009. he Shared Core of the Perfect across Englishes: A Corpus-based Analysis. In World Englishes: Problems, Properties and Prospects, eds. T. Hoffmann and L. Siebers, 309-30. Amsterdam: Benjamins.
---and L. Schäfer. 2002. he Effect of Learner Errors on POS Tag Errors during Automatic POS Tagging.
Southern African Linguistics and Applied Language Studies 20: 325-35.
Vermant, S. 1983. he English Present Perfect: A Dynamic-synchronic Approach. Wilrijk: University of Antwerpen.
Werner, V. Forthcoming. Temporal Adverbials and the Present Perfect/Past Tense Alternation. English WorldWide 34.
Wynne, T. 2000. he Present Perfect: A Corpus-based Investigation. Ph.D. diss., University of Stirling.