Structural Nativization in Postcolonial Englishes: The Complementation Profile of REMEMBER As a Case in Point

Structural nativization, that is, ‘the emergence of locally characteristic linguistic patterns and thus the genesis of a new variety of English’ (Schneider 2007: 5-6), is said to be especially prone to occur in the complementation profile of verbs (and adjectives) in Postcolonial Englishes (PCEs) (Schneider 2007: 86). However, recent research on complementation across PCEs using large corpora has focused mainly on the frequency distribution of standard patterns. This study aims to broaden the scope by conducting a corpus-based analysis using 12,000-token dataset drawn from The Corpus of Global Webbased English (GloWbE; Davies 2013) to identify possible innovative patterns in the complementation profile of REMEMBER. The varieties selected are the PCEs Indian, Bangladeshi and Sri Lankan Englishes, and British English as a reference variety. Following a holistic approach, the objective is not only to pinpoint possible instances of innovation but also explore the potential explanations of these. To do so, the complementation profile of the verb REMEMBER is considered from a diachronic perspective, mapping obsolete and regional complementation patterns and meanings (Visser 1963-1973; OED). Furthermore, the factors said to condition the linguistic evolution of PCEs are also discussed (Schneider 2007: 88-90, 99-10). The results show that there seem to be incipient traces, although very low in number, of non-standard complementation patterns in the varieties explored, both first and second language varieties. Moreover, the diachronic complementation profile of REMEMBER, plus the factors said to influence the development of PCEs, can be seen as potential explanations for most of the patterns found.


Introduction
The set of features that characterize and differentiate Postcolonial Englishes (PCEs) from first language (L1) varieties are examples of structural nativization. This term refers to 'the emergence of locally characteristic linguistic patterns and thus the genesis of a new variety of English' (Schneider 2007: 5-6). One of the areas said to be prone to undergo structural nativization is the complementation profile of verbs and adjectives, which in PCEs are claimed to allow, and at a later stage to prefer, new structures as complements (Schneider 2007: 86). However, although research on nominal verbal complementation is common (e.g. Bernaisch, Gries & Mukherjee 2014;Gries & Bernaisch 2016;Röthlisberger, Grafmiller & Szmrecsanyi 2017;Szmrecsanyi et al. 2016), clausal verbal complementation has barely been investigated at all (although cf. Bernaisch 2015;Deshors 2015;Deshors & Gries 2016;García-Castro 2019, 2020Romasanta 2017;Steger & Schneider 2012). Furthermore, these studies have usually concentrated on existing patterns (an exception being new ditransitive verbs; 1 cf. Koch & Bernaisch 2013;Mukherjee 2009) and have not explored large datasets in which new (and usually low in frequency) emerging structures acting as complements can be detected.
The present study, then, aims to go beyond establishing the frequency distribution of the traditional patterns of complementation attributed to the verb REMEMBER by exploring a 12,000-token random sample in search of these new structures. REMEMBER seems like a likely candidate to exhibit variability in its complementation profile in PCEs due to the notable changes that it has experienced since its introduction in the English language in the 14 th century. We can mention, for instance, the appearance of the gerund-participial as a new complement clause option and the complementation patterns today obsolete or regional (cf. e.g. Fanego 1996Fanego , 2007Mair 2006;Vosberg 2003). Furthermore, the sample comprises data from four different varieties of English, British English (BrE) as a reference variety, and Indian (IndE), Bangladeshi (BdE) and Sri Lankan (SLE) varieties of English as representations of PCEs, where these new patterns are expected to appear. These three PCEs are considered to be South Asian Englishes since they originated in the Indian subcontinent and, thus, they are geographically close. Consequently, they share substrate languages of the Indo-Aryan and Dravidian families and the same or a similar colonial history. Comparative studies have proved that they exhibit relevant similarities but also differences and, for that reason, that research on each individual variety is necessary (cf. Schilk, Bernaisch & Mukherjee 2012). 1 New ditransitive verbs are verbs that are not complemented with two object noun phrases in British English, but have been seen to take this basic ditransitive pattern in Indian English, this in areas where local norms have emerged (cf. Mukherjee 2009: 125). Therefore, this study will seek to identify potential explanations for the presence of the patterns found across these varieties of English. With this in mind, I will consider (i) the complementation profile of REMEMBER in present-day English, as well as its diachronic evolution, and (ii) the factors said to condition the linguistic evolution of PCEs (cf. Schneider 2007: 88-90, 99-107).
The paper is organized as follows. Section 2 provides an overview of the complementation profile of REMEMBER, both diachronically and synchronically, while Section 3 focuses on those factors likely to influence the linguistic evolution of PCEs. Section 4 sets out the methodology employed, and Section 5 discusses the findings against the theoretical background presented in Sections 2 and 3. Finally, Section 6 includes a brief summary of the main conclusions, as well as implications for the further study of new complementation patterns. 2

The Complementation Profile of REMEMBER
According to the Cambridge Dictionary (Cambridge Dictionary Online, s.v. "remember") and the Oxford Dictionaries (Oxford Dictionaries Online, s.v. "remember"), the verb REMEMBER is a polysemous word. The meanings provided in these dictionaries coincide to some extent, but not entirely. Thus, they share the following three meanings: (i) '[T]o be able to bring back a piece of information into your mind, or to keep a piece of information in your memory' (Cambridge Dictionary Online, s.v. "remember") and 'have in or be able to bring to one's mind an awareness of (someone or something from the past)' (Oxford Dictionaries Online, s.v. "remember"). With these meanings, REMEMBER can be intransitive (1) or transitive. When transitive, it can take noun phrases (NPs) (2), gerund-participial clauses (3) and interrogative clauses (4) as complement. The Cambridge Dictionary, unlike the Oxford Dictionaries, also includes declarative clauses, illustrated in (5), as a complementation option.
(1) "Where did you park the car?" "I can't remember." (Cambridge Dictionary Online, s.v. "remember") (2) I can remember people's faces, but not their names. (Cambridge Dictionary Online, s.v. "remember") (3) I remember watching my father get dressed for work when I was a kid. (Oxford Dictionaries Online, s.v. "remember") (4) Whenever a team isn't winning, people always remember who isn't playing. (Oxford Dictionaries Online, s.v. "remember") (5) She suddenly remembered (that) her keys were in her other bag. (Cambridge Dictionary Online, s.v. "remember") (ii) REMEMBER with the meaning 'to not forget to do something' (Cambridge Dictionary Online, s.v. "remember") and 'do something that one has undertaken to do or that is necessary or advisable' (Oxford Dictionaries Online, s.v. "remember") is transitive and complemented by to-infinitival complements, as in (6).
(6) Did you remember to do the shopping? (Cambridge Dictionary Online, s.v. "remember") (iii) The third meaning common to both dictionaries is 'to give a present or money to someone you love or who has provided good service to you' (Cambridge Dictionary Online, s.v. "remember") and 'bear (someone) in mind by making them a gift or making provision for them' (Oxford Dictionaries Online, s.v. "remember"). It is a transitive verb, in this case complemented by a NP, as in (7). Dictionaries Online, s.v. "remember") In addition, the Cambridge Dictionary (Cambridge Dictionary Online, s.v. "remember") lists the following meanings (not registered in the Oxford Dictionaries).

(7) [H]e has remembered the boy in his (Oxford
(iv) '[B]e remembered for sth -to be kept in people's memories because of a particular action or quality', as in (8) where the object (a NP) of REMEMBER is in subject position due to the use of a passive structure.
(8) She will be remembered for her courage (Cambridge Dictionary Online, s.v. "remember") (v) '[Y]ou remember -informal -said when you are talking to someone about something that they used to know but might have forgotten', as in (9) where REMEMBER does not take complements.
(9) We went and had tea in that little café -you remember, the one next to the bookshop. (Cambridge Dictionary Online, s.v. "remember") (vi) '[T]o hold a special ceremony to honour a past event or someone who has died'. In this case, it is a transitive verb complemented by a NP, as in (10). (10) On 11 November, the British remember those who died in the two World Wars. (Cambridge Dictionary Online, s.v. "remember") The Oxford Dictionaries (Oxford Dictionaries Online, s.v. "remember") register a number of meanings that differ from those in the Cambridge Dictionary.
(vii) 'Pray for the well-being of', as in (11), where the complement of REMEMBER is a NP in subject position due to the use of a passive structure.
(11) The congress should be remembered in our prayers.
(Oxford Dictionaries Online, s.v. "remember") (viii) REMEMBER someone meaning 'to convey greetings from one person to (another)', as in (12), where REMEMBER is complemented by a NP.
(13) She remembered herself and sat up straighter. (Oxford Dictionaries Online, s.v. "remember") (x) The verb REMEMBER is also used to emphasize the importance of what is asserted, and as such it is complemented by a declarative clause, as in (14) and (15) So, despite some differences regarding the meanings of the verb REMEMBER, the Cambridge and Oxford dictionaries agree on the fact that the verb REMEMBER takes nominal and clausal complements. In terms of clausal complementation in particular, complement clause (CC) distribution is constrained by the meaning of REMEMBER, as described in what follows.
As far as finite CCs 3 are concerned, REMEMBER can take interrogative CCs, as in (4), and declarative CCs, as in (5). Regarding non-finite complementation, REMEMBER, as a retrospective verb, belongs to a small set of verbs (together with FORGET and REGRET, among others) that allow both to-infinitival and gerund-participial CCs as complements (Fanego 1996: 71). However, verbs of this kind allow these two types of non-finite CCs in complementary distribution: gerund-participial CCs have retrospective meaning and complement REMEMBER when the meaning is 'recall' and when what is expressed in the CC 'is chronologically prior to the act of remembering' (Mair 2006: 215), as illustrated in (3); and, on the other hand, present to-infinitival CCs have prospective meaning and complement REMEMBER when it means '(take care) not to forget' (in Mair's 2006: 215 terminology), as illustrated in (6). REMEMBER with a retrospective meaning can also be followed by a perfect to-infinitival CC (Visser 1963(Visser -1976(Visser : 1876, as in (16). Despite the fact that this pattern has been described in the literature as obsolete (Mair 2006: 226;Vosberg 2003: 198), example (16) illustrates current usage in the GloWbE component, in this case from Sri Lankan English. The retrospective meaning can also be expressed by REMEMBER followed by a declarative clause, as (17) (compare with (18) and (19)).

Obsolete and regional patterns
Diachronic and synchronic research on REMEMBER has focused mainly on those clausal patterns which are still in use in present-day English, that is, the ones presented above (cf. Cuyckens, D'hoedt & Szmrecsanyi 2014;Fanego 1996;Mair 2006;Vosberg 2003). For the purpose of the present study, however, it will also be useful to consider patterns that died out or which are considered regional and/or dialectal today. The reason for this is that the input variety of the PCEs studied here is not present-day Standard English. In the case of India and Bangladesh, English entered the linguistic network in the 17 th century (Schilk 2011: 5), whereas in Sri Lanka, English arrived towards the end of the 18 th century (Bernaisch 2015: 1). This, as Brunner (2017: 25) has argued, means that 'putative innovations found in New Englishes varieties may actually be retentions from earlier forms of British English'. It is also important to bear in mind that colonizers used vernaculars rather than the standard variety. Therefore, regional and dialectal patterns may be found in PCEs too. As Mukherjee and Hoffmann (2006), Brunner (2017) and Hoffmann (2018) point out, some studies compare PCEs with present-day Standard English, which can constitute a methodological drawback if historical patterns are not taken into consideration, leading to confusion between real innovations and features retained from the original input variety. Schneider (2017: 52) notes that '[t]he superstrate input to colonial settings deserves closer consideration with respect to register differences and varying speaker styles and dialects which were brought to the colonies'. In order to try to avoid this methodological issue, I have consulted Visser (1963Visser ( -1973 and the OED (OED Online, s.v. "remember" v.1). Visser (1963Visser ( -1973 provides a broad picture of the current and obsolete clausal complementation patterns of the verb REMEMBER. He describes 'syntactical units with two verbs' (1963-1973: 2234) and provides examples of usage, as well as information as to whether or not they are still in use. The two verbs in the unit can either be in a simple catenative construction, 4 without an intervening NP between the two verbs, or in a complex catenative construction, in which an NP occurs between the two verbal constituents. This NP has 'bilateral associations, functioning as it does as the object of the first and at the same time as the subject (or passive subject) of the second verb' (Visser 1963(Visser -1973(Visser : 2234. Following Visser (1963Visser ( -1973 and OED (OED Online, s.v. "remember" v.1), I will describe the obsolete and/or regional and dialectal patterns and meanings of the verb REMEMBER.
(a) Starting with patterns in which the verb complementing REMEMBER is in infinitival form, in Middle and Modern English REMEMBER was available with the meaning 'to remind someone', which is no longer attested in present-day Standard English. With this meaning, REMEMBER could be complemented by complex to-infinitival CCs, that is, ones with an expressed subject, as in (20) and (21), and by for to-infinitival, as in (22). Today such a meaning and its associated complementation patterns are found only in regional dialects (OED Online, s.v. "remember" v.1; Visser 1963Visser -1973Visser : 2270Visser /2286 (23) and (24), although at some point in time it became a bare infinitival. This pattern was in use until at least the middle of first half of the 19 th century (Visser 1963(Visser -1973(Visser : 2315 Visser 1963Visser -1973Visser : 2315 (c) In Middle and Modern English, REMEMBER could take a complement in the past-participial form, as in (25) (Visser 1963(Visser -1973(Visser : 2397 Visser 1963Visser -1973Visser : 2397 (d) Until the beginning of the 20 th century, REMEMBER could be complemented by a pronoun plus one of the prepositions of, on or upon, as in (26), and with a pronoun plus a finite clause, as in (27) (28).
(28) I remember of seeing one or two. (1986in Dict. Amer. Regional Eng. (1996 III. 844/2; cf. OED Online, s.v. "remember" v.1) (f) Chiefly found in American English (AmE), and described as obsolete or rare by the OED (OED Online, s.v. "remember" v.1), is REMEMBER with the transitive meaning 'to cause a person to think about or recall (a thing or person)'. With this meaning, REMEMBER takes two nominal complements, as in (29). Here, the main clause is in the passive voice and thus one of the nominal complements in the active voice, the old joke, acts as subject of the passive construction, as in Harold Weschler remembered the old jokes to me. It can also take a complement clause, as in (30). This meaning could be related to meaning (a), the difference being that while in (f) what one is being reminded of is a thing or a person, in (a) one is being reminded to perform an action. In sum, REMEMBER has a complex complementation profile, which has changed and evolved, losing some complementation patterns over time, as can be seen in Table 1 below. This is relevant for the present study, since here we are dealing with varieties of English derived from an earlier input variety introduced in the new territories, which might thus differ from present-day English in many aspects, including the complementation profile of REMEMBER. These developments could explain the presence of some complementation patterns (both nominal and/or clausal) in these nativized varieties of English, which are not attested in present-day Standard English, in that some of them were available until the 20 th century in standard or vernacular forms. Hence, REMEMBER is a useful focus of study in PCEs, because these are new, developing and not yet stabilized varieties of English, potentially influenced by several external factors (language contact, substrate influence, and processes of second-language acquisition, among others, as discussed in Section 3, below). From this we might expect variability to occur to a greater degree, and to take a different form in second language (L2) varieties than in L1 varieties of English.  Visser 1963Visser -1973Visser : 2270Visser /2286 In use in regional dialects (23), (24) 'Not to forget' (Visser 1963(Visser -1973(Visser : 2315 Used until 19 th century For to-infinitival Clauses (22) 'To remind someone' (OED Online, s.v. "remember" v.1; Visser 1963Visser -1973Visser : 2270Visser /2286 In use in regional dialects Past-participial clauses (25) Not given (cf. Visser 1963Visser -1973Visser : 2397

Factors Conditioning Linguistic Evolution in PCEs
PCEs usually show particular linguistic features which are not observed in the input variety and which make these varieties different and independent with respect to the parent variety. Such linguistic features are found in a wide range of domains, from phonology to pragmatics and style, and are the result of the linguistic evolution and development of such varieties, where different factors and processes intervene (Brunner 2017: 18). As Schneider (2007: 97) explains, '[t]he linguistic processes which are involved in the genesis of PCEs are manifold and difficult to distinguish'. In this section I will briefly introduce the different linguistic processes interacting in the linguistic evolution of PCEs and its results: continuity, innovation, language contact, cultural factors, second-language acquisition processes, and typology (Schneider 2007: 88-90, 99-107). 5 Continuity accounts for the common core of the grammar of PCEs that has been preserved as a result of the direct historical transmission of the language across generations. Thus '[t]he backbone of the grammar of PCEs remains the grammar of English' (Schneider 2007: 101) with both standard and non-standard features being retained, although to different degrees depending on the type of colony. In exploitation colonies (the type of colony in which the varieties under consideration here originated), it is argued that standard patterns have been mostly preserved because English was usually transmitted through the education system (Schneider 2007: 101). However, it is important to bear in mind that part of the people who got to these colonies may have belonged to lower strata of society and may have spoken regional varieties of English, not the Standard variety. Therefore, these geographical varieties were also the input of the emergent PCEs, and features particular to these varieties persist in most PCEs (Schneider 2007: 101). In Section 2 I introduced the standard, geographical and obsolete complementation patterns of REMEMBER. Therefore, looking at when the colonization of the territories under study took place will provide an idea of what the input variety was like at the time in terms of complementation of the verb REMEMBER.
Innovation refers to 'the results of [the processes of] internal change and linguistic creativity' (Schneider 2007: 102). Here I will discuss simplification and restructuring in relation to PCEs. In the present context, simplification in PCEs is the result of internal change and linguistic creativity and involves a series of processes and mechanisms (Schneider 2007: 102). In particular, the study of PCEs has, from its early days, focused on the simplification of morphosyntactic features across these varieties (cf. Williams 1987), following the claim that PCEs tend to be structurally simpler, i.e. more explicit than the input varieties. For instance, it is claimed that PCEs prefer finite CCs (more grammatically explicit and thus less complex) than non-finite CCs (cf. Steger & Schneider 2012). Based on this generally acknowledged fact, simplification is considered to be the main explanatory factor regarding linguistic creativity in PCEs (cf. Brunner 2017; Steger & Schneider 2012; Suárez-Gómez 2017; Tamaredo 2018; Williams 1987, among others). As for restructuring, it 'involves the systematic rearrangement and reinterpretation of constituents and constituent sequences in language evolution' (Schneider 2007: 105). It results in phenomena such as 'loss of some units or rules, addition of new ones, and certainly modifications in the direction of simplification, generalization, or complexification by the addition of conditions to the application of a rule' (Mufwene 2001: 13). Restructuring may operate through reanalysis, i.e. a process whereby speakers 'analyze and understand the same constituent sequences differently in their mental grammars' (Schneider 2007: 105), which may derive in the emergence of new patterns. Schneider (2007: 106) includes grammaticalization under restructuring, in that he considers it a special case of reanalysis. Grammaticalization involves a number of processes (cf. Lass 1990: 80;Schneider 2007: 5;Trudgill 2004: 87-88), including semantic bleaching whereby a word losses lexical content meaning (Brinton & Traugott 2005: 29).
Contact is an essential factor because PCEs emerged after a process of language contact. In terms of linguistic evolution, contact contributes to change in itself, but it can also interfere with, trigger or accelerate other factors driving linguistic change in PCEs, such as simplification, restructuring and exaptation. One obvious effect of contact is the transmission of some type of linguistic material between the languages in contact. Although the most noticeable is lexical borrowing, grammar may also be affected in terms of frequency and patterns transferred from substrate languages (Schneider 2007: 107;cf. Brunner 2014cf. Brunner , 2017; Suárez-Gómez 2017, among others). Related to contact we find cultural factors, also considered relevant in the process of the linguistic evolution of PCEs, because they reflect the cultural context in which PCEs emerge. Linguistically, cultural factors may explain the emergence of lexical items and their semantics and even grammar, although it is not clear how cultural factors may affect certain levels of language. 6 Second-language acquisition processes refer to the 'universal laws of ontogenetic second-language acquisition (SLA) and phylogenetic language shift' (Schneider 2007: 89), which are often used as a potential explanation for a number of phenomena found in PCEs. Language contact and SLA processes are usually discussed jointly because PCEs are the result of language contact and SLA. Furthermore, it has been observed that cognitive constraints of second-language learning and processes of contact operating in the development of PCEs are not opposing factors, but rather complementary ones (Thomason 2001: 62). SLA processes might explain some of the shared features across PCEs which cannot be explained by language contact (cf. Mair 2003 on angloversals).
Finally, the influence of language typology has also been considered to be a determinant factor in the linguistic evolution of PCEs, since speakers 'consistently select […] forms and patterns that conform to an overarching language type, with many innovations being typologically similar in nature and strengthening one specific parameter' (Schneider 2007: 89). Although this factor is related to contact, since transfer is the process at work here, language contact goes beyond the direct borrowing and transfer of structures from one language into another.
For the present study, two of these aspects seem especially relevant: continuity, in that the complementation system is a core element of English grammar which has been transmitted and kept in PCEs; and innovation, which might also be key in explaining innovative patterns, since new structures can develop as a result of internal change and linguistic creativity, through analogical processes, for example.

Methodology
In order to survey the presence and frequency of non-canonical structures in the complementation profile of REMEMBER, I adopt a corpus-linguistic approach using the Corpus of Global Web-based English (GloWbE; cf. Davies 2013). GloWbE contains a vast amount of data from 20 different countries (including different types of varieties such as L1s, for example AmE, English as a Second Language (ESL) such as SLE, and even English as a Second Dialect (ESD) varieties, like Jamaican English), which is necessary as a means of finding incipient traces of nativization such new complementation structures. The material in the corpus was retrieved from the Internet in 2012 and can be considered to represent English as used on the web (Loureiro-Porto 2017: 455). I selected 3,000 random hits from each variety, that is, the British, Indian, Bangladeshi and Sri Lankan varieties of English.
All hits available for the verb REMEMBER were retrieved by searching for the forms <remember>, <remembers>, <remembered> and <remembering>. These four sets of samples were then randomized and the first 3,000 hits for each variety were extracted. I manually surveyed the resulting 12,000 examples and discarded those tokens which were either invalid (e.g. repeated hits) or which did not display a complement after REMEMBER (e.g. intransitive patterns). In Section 5, I will present the structures identified as possible innovative complements after REMEMBER, as well as the potential factors that may explain their emergence or presence.

Results and Discussion of Results
As noted in Section 2, the present-day complementation profile of REMEMBER includes nominal and clausal complements, which, as illustrated in Table 2 (the results of which are statistically significant at p < 0.05), are predominant in the four varieties of English. Specifically, clausal complementation is more common than nominal complementation in all sections of GloWbE. Nevertheless, of greater relevance for the present study is the fact that other patterns not described as part of the complementation of REMEMBER are also identified. Although these patterns are very low in number, their existence seems to indicate that there are indeed new structures across varieties of English. Furthermore, and as will be described below, most of them can be explained by the factors introduced in Section 3. These patterns are presented in Table 3 and will be discussed individually in what follows, providing potential explanations in each case, to then present an overall discussion. Firstly, there are instances of REMEMBER followed by several types of prepositional phrases, one which is described as an obsolete complementation pattern of REMEMBER in present-day Standard English (pattern (i); cf. OED Online, s.v. "remember" v.1; Section 2.1) and others which are not recorded in the dictionaries and grammars consulted (patterns (ii) and (iii)).
(i) REMEMBER is found followed by a prepositional of-phrase in GloWbE BD and GloWbE LK. The prepositional head of can be complemented by a noun phrase, as in (31), and by a finite clause, as in (32), and a non-finite clause, as in (33). As discussed in Section 2.1, this pattern existed historically with both the of-prepositional phrase complemented by a noun phrase and by a finite clause (as illustrated in (26) and (27) in Section 2.1). Both are now obsolete. REMEMBER with an of-prepositional phrase followed by a gerund-participial clause is still attested today in AmE, as illustrated in (28) in Section 2.1, above. Example (28) dates back to 1986, which makes it possible that its presence in L2 varieties of English is the result of continuity (Schneider 2007: 101), i.e. that it was part of the repertoire of the input variety and was retained in these varieties (cf. Section 3). On the other hand, this pattern is described by the OED (OED Online, s.v. "remember" v.1) as mainly a regional usage in the US. Thus, another possible explanation would be that the presence of this feature in other varieties of English is a result of the global influence of AmE, which is at the heart of the main economic and cultural power in the 20 th century and is considered the hub of World Englishes (Mair 2013: 261).
(28) I remember of seeing one or two. 7 (ii) An instance of REMEMBER followed by a prepositional phrase headed by as to, shown in (34), was found in GloWbE IN.
(34) When we want to become an enjoyer we do not remember as to whose things we are going to enjoy. (GloWbE IN) This pattern is not attested in the sources consulted (see Section 2.1). Therefore, the continuity factor is not likely to be responsible. Nevertheless, a potential explanation could be that there is analogy with constructions such as I was confused as to what happened or Her opinion as to what happened is unclear. The search for <remember as to whose> in GloWbE retrieves only one instance, the one presented in (34), and the search for <remember as to what> also returns just one instance. In Google, the search for <remember as to whose> produced four hits, and the search for <remember as to what> returned some 368,000. This seems to shows that REMEMBER followed by as to what could be leading such innovation, and is being followed by other types of interrogative clauses, such as those introduced by whose. So, the innovation factor might indeed explain the presence of this pattern, although such a hypothesis is tentative at this point and should be further explored in future research.
(iii) REMEMBER followed by a prepositional phrase headed by the preposition about is attested in all four varieties, as illustrated in examples (35) to (38). However, it is not accounted for in the dictionaries consulted (Cambridge Dictionary Online, s.v. "remember"; OED Online, s.v. "remember" v.1; Oxford Dictionaries Online, s.v. "remember"), in Visser (1966Visser ( -1976 or in the studies carried out on the complementation of the verb REMEMBER (cf. for example , Fanego 1996;Mair 2006). This could be an indication of analogy between REMEMBER and FORGET, which both belong to the retrospective type of verbs. In addition, analogy is cited by Schneider (2007: 103) as one of the processes through which simplification takes place in PCEs (cf. Section 3). Thus, simplification via analogy would explain the use of this construction in the PCEs under study here. Unfortunately, this hypothesis cannot be corroborated with the data available and further research is needed.
(iv) REMEMBER with the obsolete transitive meaning 'remind' and its corresponding obsolete complementation pattern (in present-day Standard English) is found only in GloWbE IN and GloWbE LK, i.e. L2 varieties, but not in GloWbE GB, i.e. the L1 variety. As illustrated in (40) and (41), REMEMBER has the meaning 'to cause a person to think about or recall (a thing or a person)', i.e. 'remind'. It is complemented by a NP and a clause in (40) and (41) (20) to (22) and (29) and (30) Visser 1963Visser -1973Visser : 2270Visser , 2286. The first of these meanings, which existed in Middle and Modern English, was 'someone is reminded to do something' and the verb REMEMBER was complemented by an NP followed by a to-infinitival clause, as in (20) and (21), and by a for toinfinitival clause, as in (22), the former being considered regional today.

(20) 'Let me remember you to do this one more kind act for me.' (21) 'Would to God that foresight had remembered me to take my cloak along.' (22) 'If a bishop… would remember him… for to keep certain moral virtues.'
The other meaning, which is chiefly found in present-day AmE, is 'someone is reminded of something'. With this meaning, the verb REMEMBER is complemented by an NP followed by a prepositional phrase, as in (29), and by a finite clause, as in (30).
(29) As the old joke goes, remembered to me by Harold Weschler, there's the faculty and the 't'aint (it ain't) faculty'. (30) I remembred to him, how often I hear him wonder [etc.] In the examples found in my data sample, both meanings are found, 'to be reminded of something', as in (40), followed by a finite clause, and 'to remind to do something', as in (41), followed by a to-infinitival. Continuity (cf. Schneider 2007: 101; Section 3) could be the factor accounting for these instances found in GloWbE IN and GloWbE LK, since they might have been part of the repertoire of the varieties of English that served as input and from which these PCEs developed. The prepositional phrase expresses recipient meaning, as do the NPs in examples (20) to (22) and the prepositional phrases in (29) and (30), and thus this pattern resembles ditransitive constructions.
Example (42) has a slightly different meaning. In this case it is 'something reminds one of something', where many [of the phones] present in the Samsung ATIVS benefits remind one to the Galaxy S (a type of smartphone). However, this pattern is not recorded as an obsolete pattern, and therefore it cannot be attributed to the continuity factor. It might, rather, be a case of linguistic creativity, a case of analogy with the verb REMIND, 8 or a case of substrate influence (cf. Section 3). However, with just one instance it is not possible to test and prove any of these hypotheses. It is interesting to compare these instances with others, such as those illustrated in (46) and (47), also taken from my data sample. Here, REMEMBER is followed by back, but back is an adjunct rather than a complement. Summing up, we can discuss these results as a whole in the following manner. In terms of distribution among varieties, we find that pattern (v) is only found in British English, patterns (i), (ii) and (iv) are restricted to PCEs (although not found in all of them as will be discussed next) and pattern (iv) is common to the four varieties under study, regardless of their status, but predominant in GloWbE IN. Focusing now on the three PCEs in detail, we can observe that out of the patterns found in this type of varieties, only pattern (iv) is present in the three varieties while pattern (i) occurs in GloWbE BD and GloWbE LK and pattern (ii) occurs only in GloWbE IN. Although the numbers are very low so as to reach a conclusion, the distribution of these patterns could point to some feature common to BdE and SLE but not shared by IndE and would confirm the claim that these varieties share some features but also diverge from each other. Further research with a larger number of instances would be needed to corroborate such claims. Considering the possible factors behind these patterns, pattern (v) found only in GloWbE GB could be a case of semantic bleaching or analogy with the phrasal verb THINK BACK. Analogy with the verb FORGET seems to also apply to pattern (iii) which is attested in all the sections of GloWbE under study. As regards the patterns found in the three PCEs, continuity could explain the occurrence of patterns (i) and (iv) while innovation (in the form of analogy, linguistic creativity or other mechanisms) could account for patterns (ii) and (iv). Finally, AmE influence could be behind pattern (i) and substrate influence is one of the many potential explanations for pattern (iv). Although these explanations remain tentative, they appear to show that not only structural nativization, but also continuity and influence from other languages or varieties should be considered regarding these innovative patterns.

Conclusion
In this paper I set out to explore the possible manifestation of structural nativization in the complementation profile of the verb REMEMBER in PCEs. I have done so by performing a preliminary study using a random sample with data from four components of GloWbE, which represent British English and three PCEs, Indian, Bangladeshi and Sri Lankan Englishes. Following a holistic approach, I have identified new structures and their frequency, and have tried to provide an explanation for their presence, taking into account the diachronic evolution of the complementation profile of the verb REMEMBER and the factors said to condition the linguistic evolution of PCEs.
This pilot study shows that patterns exist across the four varieties of English surveyed that are not described as part of present-day Standard English, even in the L1 variety. Despite the fact that these patterns are anecdotal in number, their interest lies in that they might represent incipient traces of structural nativization and/or remnants from the original input variety, and that they might be caused by several factors pertaining to the linguistic evolution of PCEs. These factors are, among others, continuity, innovation, and the influence of American English. Furthermore, GloWbE GB, an L1 variety, also shows an innovative pattern (remember back). Thus, although not a PCE, innovation also seems to be at work here, and this pattern could be the result of analogy and/or semantic bleaching. Despite the fact that the low number of hits make such hypotheses highly tentative, the findings seem to corroborate claims that complementation in general, and the verb REMEMBER in particular, are a prolific area of research in which PCEs are likely to develop structural differences from the input variety. Moreover, it does not seem wholly accurate to apply the labels new and innovative to these patterns in general. While some may be considered the result of structural nativization and thus labelled innovative patterns, others could be the result of the continuity factor, since they formed part of an earlier stage of the English language, and thus would in fact be old patterns.
In sum, this study appears to corroborate the claim that structural nativization is likely to manifest itself in the complementation profile of verbs. Moreover, we have seen how important it is to consider the diachronic evolution of the phenomenon under study, since earlier versions of English may explain features that have been retained in PCEs but which are not present in L1 varieties of English. Other factors concerning the linguistic evolution of PCEs are also key to understand the origin of these patterns. However, some of these, such as language contact and substrate influence, are not thoroughly explored here. Therefore, further research should focus on substrate languages to see whether some of these patterns may correspond to existing structures in substrate languages that have been transferred to English. Furthermore, with such a low number of tokens it is clear that future research will need (i) to resort to larger datasets, due to the low frequency of these non-canonical structures, and (ii) to test whether the patterns identified are found in other varieties of English, and if the potential explanations proposed here are valid.
BES-2015-071543) and the University of Vigo. I would also like to thank the anonymous reviewer who provided helpful comments on earlier drafts of the manuscript.