2.1. The expansion of prevent+ -ing without from in present-day English

This section surveys the current usage of prevent, focusing on the competition between -ing complements with and without from in British English. Developing Mair’s (Reference Mair, Carls and Lucko1999) research on the absence of from in complementation, Mair (Reference Mair2002, 112–115) analyzes data from four parallel corpora: LOB (British English, 1961), FLOB (British English, 1991), Brown (American English, 1961), and Frown (American English, 1992).Footnote ⁴ He demonstrates that from-less -ing complements have been increasing in contemporary British English. The data in Table 1 are drawn from Mair (Reference Mair2002, 112):.

Table 1.

The complementation of prevent in British English (LOB, FLOB) and American English (Brown, Frown)

Table 1 shows that the preposition from is almost consistently present in American English, a trend that remained steady from the 1960s to the 1990s. In British English, by contrast, from-less -ing constructions were already widely used in the 1960s, and their frequency increased over time. By the 1990s, when constructions with and without from occurred with equal frequency in British English, the gap between British and American English had become even more pronounced. Thus, the findings suggest a rise in from-less -ing in recent British English

The same applies to Tajima (Reference Tajima1995, 10–11), whose work, written in Japanese, is less frequently cited in the previous literature. He analyzes the KB Corpus (Konomi-British-English Corpus) and the KA Corpus (Konomi-American-English Corpus)Footnote ⁵, which contain texts from the 1980s onwards, and shows that there are 30 examples with from vs. 14 examples without from in British English, and 34 examples with from vs. 3 examples without from in American English. From-less -ing complements in KB are clearly more common than in LOB in Mair’s statistics, though less common than in FLOB. The lower frequency of from-less complements in KB compared to FLOB may be attributable to the slight difference in the dates of the corpora. Alternatively, the overall proportions of these constructions may vary considerably depending on the datasets examined.

There are also other studies that suggest a significant rise in from-less -ing constructions in British English. Sellgren’s (Reference Sellgren, Lenker, Huber and Mailhammer2010, 53) analysis of the British National Corpus, which contains data from the late 20^th century, identifies 128 examples of from -ing compared to 113 examples of from-less -ing.Footnote ⁶ Rudanko (Reference Rudanko, Leistyna and Meyer2003, 280) also supports this trend. Although he does not specifically address the expansion of from-less -ing constructions, he notes that -ing constructions with and without from are ‘approximately of the same frequency’ in the spoken British English component of the Bank of English (a monitor corpus) compiled by COBUILDFootnote ⁷, suggesting how fully established from-less -ing constructions were in British English around the turn of the 21^st century.

There are, however, some studies that present slightly contradictory results, although they all show a fairly frequent use of from-less constructions in British English. Mair and Leech (Reference Mair, Leech, Aarts, McMahon and Hinrichs2021, 266–267), updating Mair (Reference Mair2002), investigate British English 2006, which contains data from 2003 to 2008Footnote ⁸, and finds 23 examples of from -ing and 15 examples of from-less -ing. The from-less variant is still noticeably frequent, but not as frequent as the from -ing variant. Likewise, Heyvaert et al. (Reference Heyvaert, Rogiers and Vermeylen2005, 84–85) present the statistics shown in Table 2, drawing on data from earlier studies and comparing them with their own data extracted from two components (Times and UKspoken) of COBUILD.

Table 2.

Prevent and three gerundial constructions in previous studiesFootnote ⁹

As their methodology differs significantly from that of the studies mentioned above, the results in Table 2 are not directly comparable with those in other studies. First, they are mainly concerned with the relationship between the possessive and objective cases in the subject of -ing forms in from-less constructions, rather than with the contrast between from -ing and from-less -ing. Second, they examine only those examples in which the subject of -ing is a personal pronoun. Still, it is probably worth noting that their recent data, extracted from the COBUILD Corpus, yield a notably high proportion of from -ing complements. Hence, it will be interesting to conduct further research into how extensively from-less constructions are employed around the turn of the 21^st century, which is the aim of the present study.

Incidentally, the sharp decline in the ‘possessive subject + -ing’ construction around the turn of the 20^th century, as shown in Table 2, is also intriguing. Although the possessive type is now very rare in contemporary English, this variant will still be taken into account in the following discussion. The data in Aarts et al. (Reference Aarts, Wallis, Bowie, Seoane, Fariña and Palacios–Martínez2018, 152–153) drawn from the London-Lund Corpus and the British Component of the International Corpus of English show that the possessive type accounts for only about 4% of all possessive/objective + -ing constructions.Footnote ¹⁰ Moreover, the possessive type is generally considered more formal in contemporary English (cf. Quirk et al. Reference Quirk, Greenbaum, Leech and Svartvik1985, 1063–1064; Huddleston and Pullum Reference Huddleston and Pullum2002, 1192).

2.2. Data

The primary dataset in this study is drawn from two relatively casual contemporary newspapers: The Daily Mail (British) and USA Today (American). Using the electronic database Lexis+ Footnote ¹¹, I extracted all forms of prevent from all issues published between October and December 2020, obtaining 912 examples from The Daily Mail and 267 examples from USA Today, after removing clear duplicates. It is interesting to note that prevent occurs significantly more frequently in The Daily Mail than in USA Today, which may be attributable to various factors, such as differences in overall word counts, subject matter, and other variables. However, I leave this issue outside the scope of the present research, partly because the focus is on variation in the complementation of prevent, and partly because Lexis+ is a database rather than a corpus, meaning that the full text is not accessible to users. Therefore, standard corpus linguistic methodologies based on word counts cannot be applied. Nevertheless, it remains a powerful resource for linguistic analysis.

In addition to the main dataset, I will also examine the British Academic Written English (BAWE) corpus as a supplementary resource. The corpus was released in 2008 and comprises data from 2000 to 2007. It consists of academic assignments written by students, including native and non-native speakers of English, and ‘was developed at the Universities of Warwick, Reading and Oxford Brookes under the directorship of Hilary Nesi and Sheena Gardner (formerly of the Centre for Applied Linguistics, Warwick), Paul Thompson (formerly of the Department of Applied Linguistics, Reading) and Paul Wickens (School of Education, Oxford Brookes), with funding from the ESRC (RES-000-23-0800)’ (https://warwick.ac.uk/fac/soc/al/research/collections/bawe/how_to_cite_bawe/).Footnote ¹² The use of this corpus is intended to substantiate the discussion through the examination of unpublished, and therefore largely unedited, texts. I will extract examples of prevent only from the assignments written by native speakers in this corpus and use them in the following analysis (details in Section 4). Most of the writers in this corpus were born in the 1990s. Assuming that they were around 20 years old when they drafted the assignments, the data obtained from this corpus provide a useful perspective on the usage of prevent in a period largely comparable to that represented by the data from The Daily Mail and USA Today (December 2020), since the generation that contributed to BAWE is likely to have been actively involved in social interactions in 2020. In other words, the generation that contributed data to BAWE and the generation represented in the two newspapers should at least partially overlap. It will be interesting to observe how the usage of prevent differs across datasets of different types.

2.3. Relevant examples

This section defines the criteria for the examples examined in this study. First, all instances of the verb prevent are considered relevant as long as they include complements. Hence, both finite (1) and non-finite (2) forms of prevent are counted as examples:

1. (1) A decent career has incorporated 42 England caps and a £16million move to Arsenal, but injuries prevented it from becoming great and Brighton was his best option in October after being released by Watford. (The Daily Mail, 23 Nov 2020)

2. (2) It is 70 per cent effective at preventing people falling ill with Covid, initial results suggest. (The Daily Mail, 9 Dec 2020)

Second, the main part of the discussion focuses on -ing forms used as complements of prevent. Other examples, such as the following, in which the verb governs a noun phrase, are excluded from the analysis:

1. (3) The U.S. Food and Drug Administration authorized the first vaccine to prevent the disease Friday night. (USA Today, 14 Dec 2020)

Examples in which -ing is used as a noun are also excluded. See, for example:

1. (4) Alexa also shares tips to prevent the spreading of the virus. (USA Today, 3 Nov 2020)

Third, I restrict the discussion to those examples in which the subject of -ing is explicit, since this is the context in which the alternation between -ing constructions with and without from occurs. Therefore, examples such as (5) are excluded, whereas (6) and (7) are included:

1. (5) . . . as earnings surged even though Government interventions have prevented sweeping insolvencies. (The Daily Mail, 19 Nov 2020)

2. (6) The craft were used to form a barrier to prevent them from travelling upstream. (The Daily Mail, 2 Oct 2020)

3. (7) . . . and Twitter users were prevented from liking or replying to the tweet. (USA Today, 4 Nov 2020)

The subject of an -ing form can appear in the possessive or objective case in the active voice, whereas in the passive voice it becomes the grammatical subject and therefore takes the nominative case.

Finally, the unedited texts in BAWE contain several examples of to-infinitival complements, which serve as an alternative to the -ing construction. In addition, they include an example of a that-clause functioning as the complement of prevent. These examples are excluded from the analysis of -ing forms, although they are discussed as separate but related constructions in Section 4.

3.1. Prevent in the active voice

Of the 912 examples of prevent in The Daily Mail, 464 occur with -ing complements that meet the conditions stated in Section 2.3. Likewise, among the 267 examples of the same verb in USA Today, 96 are relevant. Furthermore, most of the relevant examples are found in the active voice, namely 437 in The Daily Mail and 94 in USA Today, which will be examined in this section. As expected, the frequent occurrence of from-less -ing is a notable feature of British English. Table 3 presents the raw frequencies of -ing complements with and without from in the two newspaper datasets:

Table 3.

Prevent + (from) -ing in The Daily Mail and USA Today (raw frequencies)

All relevant examples in USA Today exhibit the from -ing construction, whereas The Daily Mail makes extensive use of from-less examples, which are in fact more frequent than from -ing ones. The use of from-less examples in British English appears to have advanced even further in recent years than in the data examined in previous studies (cf. Section 2.1). It is therefore likely that the use of from-less -ing is continuing to expand in contemporary British English, at least in relatively casual styles such as that found in The Daily Mail.

In this construction, the subject of -ing appears in either the possessive or the objective case. The data presented by Heyvaert et al. (Reference Heyvaert, Rogiers and Vermeylen2005, 84) (see Section 2.1) show a clear rise in objective subjects as opposed to possessive ones from the 18^th to the 20^th century. In the present study, I will examine personal pronouns used as subjects of -ing separately from other noun phrases. This is because the possessive use of the subject of -ing in general (not specifically restricted to its use with prevent) is more frequent with pronouns than with other noun phrases (Quirk et al. Reference Quirk, Greenbaum, Leech and Svartvik1985, 1063–1064).Footnote ¹³ The Daily Mail and USA Today provide a single example each in which the subject of -ing consists of both a personal pronoun and a noun (see [8]), which I include in the category of other noun phrases:

1. (8) . . . and prevented me and my lead scout from getting serious injuries or death. (The Daily Mail, 10 Nov 2020)

When the subject of -ing is her as in (9), it is practically impossible to determine whether it is possessive or objective. Such examples are treated separately.

1. (9) . . . but a local court prevented her seizing it. (The Daily Mail, 29 Nov 2020)

See Figure 1, which presents the raw frequencies of all relevant types of -ing constructions:

Figure 1.

Occurrences of prevent with -ing in the active voice in The Daily Mail and USA Today (raw frequencies).

This graph is principally intended to illustrate the structural variation in British English, since USA Today consistently uses the from -ing construction regardless of whether or not the object of prevent, which also serves as the subject of -ing, is a personal pronoun. In The Daily Mail, constructions are more varied, but as expected (cf. Section 2.1), possessive subjects of -ing are nearly extinct. In other words, the competing constructions in the datasets under examination are now limited to from-less -ing with objective subjects and the from -ing construction. There appears to be a notable contrast between personal pronouns and other noun phrases. With personal pronouns, the construction with from remains dominant, whereas with other noun phrases, the from-less construction is more prevalent. The recent increase in from-less -ing constructions thus seems to be driven primarily by noun phrases other than personal pronouns.

3.2. Prevent in the passive voice

Relevant examples of prevent in the passive voice are very infrequent: The Daily Mail provides 27 examples, while USA Today contains only two. In the passive, there is a strong tendency for -ing to be preceded by from, as illustrated in (10) and (11):

1. (10) They are also prevented from slamming doors . . . (The Daily Mail, 13 Sep 2020)

2. (11) By contrast I, aged 69, am prevented by law from taking the very minor risks with my life or yours . . . (The Daily Mail, 8 Nov 2020)

In the dataset used in this study, all relevant examples exhibit the from -ing construction, irrespective of whether the subject is a personal pronoun or a noun phrase of a different type.

Although the passive use of prevent itself is very infrequent, it is probably safe to conclude that the tendency to use the from -ing construction is firmly established even in British English. Rohdenburg (Reference Rohdenburg1995, 86) identifies three passive examples in LOB, all of which show the from -ing construction. Likewise, Rudanko (Reference Rudanko, Leistyna and Meyer2003, 281) investigates rare examples of the passive use of prevent in the Spoken British English section of the Bank of English and observes that ‘they are quite regularly of the from -ing type’. See also Huddleston and Pullum (Reference Huddleston and Pullum2002, 1238), who state: ‘Passivisation is restricted to the prepositional construction’. The results of the present study align with these previous findings. Thus, the complementation of this verb in the passive voice appears to be independent of the ongoing change in its active usage.

4.1. Prevent in the active voice

Since the constructions of prevent are notably varied in British English, I have decided to examine an additional British English dataset to gain a clearer understanding of this phenomenon. BAWE serves this purpose. I have extracted only the documents written by native speakers from BAWE, which I refer to as BAWE (L1) hereafter. This dataset differs from The Daily Mail and USA Today, as it consists of writings by students in academic settings. In other words, BAWE (L1) represents unpublished and relatively unedited language use in contemporary British English, although the assignments in the dataset are generally of high quality, having received grades of merit or distinction. Chronologically, it is broadly comparable to The Daily Mail (cf. Section 2.2).Footnote ¹⁴

BAWE (L1) provides a total of 327 relevant examples in the active voice, none of which involve the possessive subject plus -ing construction. Ambiguous examples with her as the subject of -ing are also absent. Thus, in this corpus, the structural competition in the active voice is essentially between from -ing and from-less -ing, insofar as -ing complements are concerned, as in:

1. (12) Limited space prevents me from providing an exhaustive study of each and every one of the rituals employed . . . (BAWE)

2. (13) In order to prevent boundary effects distorting the model, the main box was surrounded by 26 ‘ghost’ boxes. (BAWE)

Here again, there appear to be differing tendencies depending on whether the subject of -ing is a personal pronoun or not, as shown in Figure 2:

Figure 2.

Occurrences of prevent with -ing in the active voice in BAWE (L1) (raw frequencies).

The from-less construction is more extensively used when the subject of -ing is a noun phrase other than a personal pronoun, as in The Daily Mail (see Section 3.1). However, the extent to which from-less -ing occurs differs noticeably between The Daily Mail and BAWE (L1): from-less examples are less frequent in BAWE (L1) than in The Daily Mail.

It is not immediately possible to determine the cause of this difference. The difference in dates may be one factor. As mentioned in Section 2.2, the authors represented in BAWE and those in The Daily Mail (2020) are indeed largely comparable, but it is also true that BAWE (L1) predates the dataset of The Daily Mail used in this study. However, from-less examples occur even less frequently in BAWE (L1) than in Mair’s (Reference Mair2002) data from the 1990s, which are earlier. Thus, the date of BAWE (L1) does not necessarily explain this difference in tendency.

Stylistic factors may play a role: the expansion of from-less constructions may be more prevalent in printed materials than in academic assignments. If so, this may be due either to a probable prescriptive sense in academic writing – a factor in favour of from -ing – or to a possible preference for brevity in newspapers – a factor in favour of from-less -ing. Alternatively, it could be attributed to the relatively informal nature of the tabloid The Daily Mail.

Another relevant factor may be that newspaper English tends to be ‘receptive to (and in a good many instances also productive of) innovations or changes “from below”’ (Hundt and Mair Reference Mair, Carls and Lucko1999, 236).Footnote ¹⁵ On the other hand, the absence of possessive subjects of -ing in BAWE (L1) may represent a more advanced stage than in The Daily Mail, since the possessive form, as opposed to the objective form, is generally more archaic (cf. Section 2.1).

Finally, BAWE includes a small number of rare constructions that do not involve -ing and are therefore excluded from the above analysis. The most striking is example (14), in which prevent is followed by a finite clause introduced by the conjunction that. This instance comes from a third-year student whose education was entirely in the UK.

1. (14) . . . and therefore more likely to take action to ensure that she does or prevent that she doesn’t. (BAWE)

Obviously, this usage appears to have been influenced by ensure that she does, which directly precedes and parallels prevent that she doesn’t. Still, the use of a that-clause is noteworthy, as it represents an older complement form with this verb – one that was subsequently replaced by the to-infinitive and later by the -ing construction (cf. Section 5). This construction is, however, not wholly obsolete. The Oxford English Dictionary (OED) notes that prevent occurs ‘occasionally with that-clause,’ although -ing constructions are more usual (s.v. prevent, II.9.b).Footnote ¹⁶

The second type is more common and therefore not necessarily exceptional. Prevent is also followed by infinitival complements in BAWE (L1), and this occurs repeatedly. The dataset provides seven examples of this type in the active voice, as in:

1. (15) RV helps to maintain alveolar patency by preventing all the air to be expelled . . . (BAWE)

This is interesting, since infinitival constructions are not recognized in usage guides for contemporary EnglishFootnote ¹⁷, although they are attested in historical data (cf. Section 5). All seven examples show a subject of -ing other than a personal pronoun. Hence, structural variation appears to be greater when subjects other than personal pronouns are involved. Furthermore, overall complementation patterns exhibit greater variation in the relatively unedited texts of BAWE (L1) than in The Daily Mail, although examples of possessive subjects are absent in the former.

4.2. Prevent in the passive

Examples of the passive voice are very infrequent and, as in The Daily Mail, the structural patterns of the passive are less varied. BAWE (L1) provides a total of 19 examples with -ing, all of which include the preposition from, as illustrated in the following example:

1. (16) under which a contract was prevented from conferring rights upon anyone except the parties to it. (BAWE)

Interestingly, one of the 19 examples features both from -ing and to-infinitival complements:

1. (17) Women providing capital for investment were prevented from acting as representatives or to carry out long-distance trade themselves, removing their ability to ensure the safety of their returns. (BAWE)

Apart from these 19 examples, there is one example in which prevent directly takes a to-infinitive:

1. (18) if they are prevented to do so the future survival of genetically diverse groups may be put placed in jeopardy. (BAWE)

Infinitival complements are rare with prevent, whether in the active or passive voice, and are generally not recognized in usage guides or dictionaries of contemporary English. Whether the construction found in BAWE (L1) represents a sporadic resurgence of an older usage or reflects the influence of the infinitival construction associated with other verbs such as forbid remains a matter of debate. Further discussion is provided in the following section.