Skip to navigation – Site map

On the polysemy of help

Mathilde Pinson


The verb help is highly polysemous and some of its meanings even seem contradictory. In order to try and understand how this polysemy came into being, I propose a look back at examples from the early modern period onwards, so as to trace the link that unites each secondary meaning with the original meaning of help. These observations are compiled in a synthetical table and a graph, which show how a limited range of semantic features and syntactic structures interact to generate meaning.

Top of page

Full text


1Looking at lexical items from a speaker-oriented perspective can be done from a variety of angles. One of the first speaker-dependent aspects which have been revealed in the literature is the influence on language of the speaker’s temporal and/or spatial localisation, i.e. deixis, e.g. Benvéniste [1966]; Maingueneau [1976]. Others also focus on how lexical items may reflect the subjectivity of the speaker, in terms of axiological or epistemic evaluation, e.g. Kerbrat-Orrechioni [2002 (1980)]. Contrasting with formal frameworks, which posit the autonomy of syntax, most speaker-oriented approaches study how meaning emerges in context, through an intricate combination of both syntax and semantics. One of the major breakthroughs of speaker-oriented approaches is to have shown that the grammatical well-formedness of a sentence does not guarantee its acceptability [Culioli 1990-1999]. Many perfectly correct sentences cannot be uttered because they would violate other types of constraints related to pragmatics and subjectivity. Indeed, the focus is on language as an act of communication and utterances have to be apprehended in context, notably in interaction. This new direction is accompanied by a thorough study of semantics and of the subtle differences in meaning that distinguish alternative variants with quasi-similar informational content. In order to go deeper into the semantics of lexical and grammatical items, and to understand underlying mechanisms, speaker-oriented researchers often endeavour to identify their invariant pole.

2Apart from deixis, which is unrelated to the case of the verb help, all the other dimensions mentioned above are relevant to the study of this verb. More specifically, this paper will focus on the polysemy of the verb help, the emergence of meaning in context and particularly the emergence of new senses over the course of history. The main research questions are: 1) what do the various uses of help have in common? and 2) how can we account for the appearance of the secondary uses of help?

3I will first address the semantics of prototypical help, before focusing on the emergence of each of the other meanings of help. I will finally propose a semantic grid and a graph for the verb help, in order to summarise what its various meanings share semantically and how they diverge.

1. The semantics of prototypical help

4The notion of prototype [Rosch 1973] refers to the ‘best’ representative of a category. The prototype theory makes it possible to apprehend categories in all their complexity, as it includes the possibility for members of a given category to be more or less central, therefore challenging the conception that all members have to share the same set of features.

5In that respect, the prototype theory converges to some extent with Culioli’s notional domains, with their borders, their interior and exterior and their organising centres [Pauly 2010: 86].

6When native speakers of English are asked about the meaning of the verb help, they invariably associate it with the notion of assistance. Moreover, a sample of 100 occurrences of the verb help in the Corpus of Contemporary American English [Davies 2008-] exhibits this meaning of assistance 95 % of the time. This prototypical meaning is also the oldest meaning of the verb, as the Oxford English Dictionary testifies (s.v. help).

7As a dynamic verb (in the sense of Talmy [2000] and Jackendoff [1990: 133]), the central meaning of the verb help implies an interaction between protagonists followed by an effect. Prototypically, it associates an animate helpee with an animate helper, who performs some task for the former.

1.1. Participation of each protagonist

8The exact share of action undertaken by each participant is, however, not explicitly encoded in the verb. In some cases, both participants take part in the action in (relatively) equal amount. For instance in the following example, the process carry it up to my room, which refers to the lifting of a heavy referent (a freezer), implies that the action is likely to be performed by both participants together:

(1) “I have to keep this freezer here,” Thorn said.
“You can’t put it there. It’s not convenient.”
“Will you
help me carry it up to my room?” (COCA, 2011, FICTION)

9Conversely, in other cases, the type of infinitive verb following help indicates that the helpee is the one who performs the process encoded in the infinitive clause, not the helper. This is what we observe in the case of predicates that express individual physiological or psychological processes, such as the following:

(2) Think only of your own welfare and mine and do the right thing, the Christian thing, and help me to die properly. (COCA, 1994, FICTION)

(3) Mr. Griffith helped me to understand who I could become if I kept trying. (COCA, 1998, SPOKEN)

10In the case of these internal processes, the helper performs some task towards an outcome which concerns the helpee alone. These examples contrast with cases in which the part played by the helpee is insignificant. In an utterance like:

(4) Voice-over: Things were so bad, they concluded the only way to end their suffering was to commit suicide. After a 43-day hunger strike, her mother, brother and sister all slit their wrists. But not even that worked […].
Oufkir: After 43 days of a hunger strike, you cut your veins but the blood doesn’t flow. He
helped me open it with a piece of broken glass. (COCA, 2001, SPOKEN)

the helpee may play no part at all in the opening process. It is quite clear that she wants the infinitive process to happen, but she probably does not do anything towards this aim, since such an action is likely to be performed by a single agent.

The exact share of action undertaken by each participant is therefore not what defines ‘help’. The type of complementation, however, has been used to specify the role of each participant, as the question of the presence versus absence of to after help suggests.

1.2. To versus zero after help: an overview

  • 1 It is difficult to be very specific regarding the date when this change occurred. Indeed, it is sti (...)

11In Old English, the verb help was associated with an animate recipient in the dative, optionally followed by an adverbial clause denoting purpose (Los [2005]). It subsequently became a ditransitive verb, as the possible wh-extraction of (a part of) the object suggests1 (cf. Chomsky [1980]):

(5) He helped [you] [to build a house].
What did he help you to build?

(6) You need his help [(in order) to build a house].
(6’) *
What do you need his help (in order) to build?

12The purposive origin of the infinitive clause explains why to was initially present, but its use has now ceased to be compulsory. While formal frameworks posit that the infinitive marker is completely devoid of meaning [Chomsky 1957: 100], many researchers working within the framework of speaker-oriented theories have tried to identify a semantic invariant for to (e.g. Adamczewski [1982]; Cotte [1982]; Chuquet [1986]), and so have some cognitivist researchers (e.g. Egan [2008]).

13Among the speaker-oriented linguists who have looked for a core value of to that could subsume both the preposition and the infinitive marker, the work of Duffley [1992], in particular, has proved seminal. Regarding the alternation between to and zero after help, he claims that the absence of to is correlated with more agentivity on the part of the helper.

14This implies that inanimate subjects are incompatible with the omission of to, and so are infinitive predicates denoting individual psychological or physiological processes (cf. supra). It also implies that the speaker can express the directness or indirectness of the helping process through the omission or the use of to (cf. Dixon [1991]; Fischer [1995]). The well-foundedness of these observations can be illustrated by the following examples, dated from the 19th century and the early 20th century:

– animate vs. inanimate subject:

(7) I’ll help you Ø catch the Englishman. (James Fennimore Cooper, The Pilot, 1824)

(8) Here is a ring whose signet may help us to understand each other. (James Fennimore Cooper, The Bravo, 1831)

– individual process:

(9) And you’d be doing her a much more neighborly act if you helped her to realize the value of money, instead of encouraging her to throw it away. (COHA, 1928, FICTION)

– direct help:

(10) However, he told her that she would have to eat up all the meat. She said she couldn’t. “I will help you Ø eat,” said the man. (The Journal of American Folklore, 1888)

– indirect help:

(11) [She] could neither sit nor be made to sit; she had an attendant to move her and help her to eat. (Publications of the Oriental Translation Fund, University of Michigan, 1922)

15As can be seen in (10) and (11), when it comes to eating, the helping process can be done by eating a part of the food (direct help) or by, say, holding a fork for the eater (indirect help).

16However, the clear-cut semantic distinction between to and zero had considerably loosened by the mid-twentieth century. Corpora show that the omission of to has now become compatible with all of the above cases. It can occur with inanimate subjects and indirect help:

(12) Pepsodent Tooth Paste helps Ø stop decay before it starts. (Saturday Evening Post, 1950)

as well as with individual psychological or physiological processes:

(13) Mr Parish helped me Ø understand this problem when I several times proposed moving a painting or a statue to take a photograph. (COHA, 1958, MAGAZINE)

17This innovation is undoubtedly partly due to reasons of economy, but other possible causes can be suggested. None of these factors constitute categorical constraints but they can be hypothesised to have had some influence on the increase of the use of zero. They include euphonic and cognitive principles such as stress alternation [Schlüter 2003] e.g. He ˈhelped me ˈmake my ˈbed. Another such potential factor is horror aequi (see Bolinger [1979]; Rohdenburg [2003]; McEnery & Xiao [2005]) e.g. ?I want to help him to write (see Pinson [2009: 159-163] for a quantitative analysis of the influence of horror aequi on the omission of to after help in the 18-19th centuries).

18The semantic distinction between the two variants may also have contributed to the decrease in the use of to, because of the influence of advertising language on everyday English. Indeed, as example (12) illustrates, advertisers generally favour the zero version, because it is an excellent way to subtly enhance the efficiency of the product advertised, while remaining within the bounds of legalities of truth (Pinson [2009: 189-191]). Because of increasing exposure to the zero version, speakers/hearers ceased to conceive it as encoding direct help and started to see it as the default version, a process known as habituation Haiman [1994]. So, paradoxically, the exploitation of the semantic distinction may therefore have led to its gradual demise.

19The omission of to is on its way to becoming the statistical norm, except in syntactically and cognitively complex co-texts, such as the negation of the infinitive verb, and the passive voice (cf. Wierzbicka [1988: 49]). Between 1961 and 1991, the percentage of the use of zero after help increased from 22 % to 60.8 % in British English and from 68.3 % to 81.9 % in American English [McEnery & Xiao [2005]. This sharp increase suggests that the semantic distinction is no longer a meaningful criterion to explain the alternation between to and zero.

20Recent research (e.g. Egan [2008], who works within a cognitive framework) now tends to acknowledge that “the semantics of the two constructions are […] very similar [in the case of help] and it should not therefore be surprising that in many instances speakers may choose to use either of them to encode one and the same situation” (Egan [2008: 214]) and that “this distinction between the two construction types is blurred in the case of help” (Egan [2008: 243]).

2. The core value of help

21As suggested by the fact that the NP following help was originally a dative, the semantic core of help seems, at first sight, to be based on the fact that the effect of the process is seen as beneficial for the helpee. When a helper helps a helpee, it is not clear who does what exactly, and in which proportion, but what is certain is that the result is seen as positive for the helpee.

22In that respect, help is an evaluative verb that constitutes an indication of the subjectivity of the speaker. Based on Kerbrat-Orecchioni’s typology of verbs [2002 (1980): 113 ff.], help can be said to belong to the list of inherently subjective verbs. The evaluation emanates from the speaker; the object of the evaluation is the process expressed in the infinitive clause and the type of evaluation is axiological. More specifically, the speaker evaluates that the process expressed by the infinitive is good for the referent of the first object of the verb.

2.1. The polysemy of help

23The prototypical semantics of help reflects a process whereby some participant acts in a way that is beneficial to some other participant, but does the euphoric denotation associated with prototypical help apply to the other meanings of this verb?

  • 2 The near-synonyms given here are used to illustrate the different meanings of help. They are not ne (...)

24Although dictionaries do not reflect the complexity of polysemy, they can constitute a useful starting point for a first tentative analysis of help, which will be refined in the remainder of this paper. Dictionaries usually list five different meanings for the verb help. Apart from the prototypical meaning (15a and 15b), one notes the meanings “alleviate/cure” (16); “contribute” (17); “provide/serve” (18); as well as the meaning “not be able to control or stop”, associated with the phrase cannot help (19)2.

25Each of these senses is illustrated in the following utterances:

(15)  (a) They helped me to get to where I’m at today. (COCA, 2012, NEWS)
       (b) One of her older brothers took her hand to
help her across the highway. (COCA, 1998, FICTION)

(16) Amma applied a paste of turmeric and egg yolk to help the pain and swelling. (COCA, 2000, FICTION)

(17) The Government’s decision to axe half-price coach fares for the over-60s saw one million fewer pensioners travel with National Express last year, helping to send the transport firm’s annual pre-tax profit down 9 % to £164.1 million. (Evening Standard, 28/02/2013)

(18) He helps himself to a melon from a bowl. (COCA, 1997, FICTION)

(19) This woman can’t help but giggle. (COCA, 2012, SPOKEN)

26Quite clearly, the beneficial value is absent from sentence (17), in which help is closer to the meaning of the verbs cause or contribute (cf. Mair [1995], [2002: 126]; Deléchelle [1998]). By contrast, the verb help in sentence (19) corresponds to notions such as refrain or avoid. How is it possible to reconcile such contradictory meanings as ‘contribute’ and ‘avoid’?

2.2. A semantic invariant?

27Researchers working within the framework of speaker-oriented theories or psychomechanics usually endeavour to find an invariant semantic value for polysemous words. This is especially true for grammatical items (e.g. Joly [1980]; Adamczewski [1982]; Chuquet [1986]; Culioli [1990-1999]; Franckel & Paillard [2007]), but also for lexical items (e.g. Picoche [1995]; Victorri & Fuchs [1996]; Cadiot [1994]; Franckel & Lebaud [1995]; Paillard [2000]; de Vogué [2004], Pauly [2010]). For instance, Picoche, following Guillaume, posits that polysemous lexical items have a ‘potential meaning’ which can account for the multiple senses observable in context; Franckel and others, following Culioli, endeavour to identify the ‘schematic form’, or abstract scenario of polysemous verbs.

28However, the existence of a semantic invariant is far from being straightforward, as shown notably by Girard [1996], Kleiber [1999], Puckica [2008] and Toupin & Lowry [2010]. Approaches in favour of an invariant pole claim that the use of the same word for different meanings cannot be attributed to chance and that there has to be some semantic feature in common between all the senses of this word (except of course in cases of homonymy). Although it is evident that polysemy is not random, it is far from certain that all the meanings of a verb share a common semantic core or display a common scenario.

29When Paillard, Franckel, and others, work on schematic forms, they manage to account for the various literal and figurative meanings of verbs, but seldom do they have to describe such strikingly contradictory meanings as those observed in the case of help. In order to reconcile all the meanings of help, and in particular ‘avoid’ and ‘contribute’, one would have to bridge an unbridgeable gap. The only type of scenario that could reflect the meaning of help would be that the verb help means that an action a is related to a situation x. Is it really appropriate to reduce the meaning of help to such a broad scenario, which could also correspond to hundreds of other verbs?

30As Lakoff [1987: 95-96] suggests, “categories on the whole need not be defined by common properties” and “central members are linked to other members, which are linked to other members, and so on”. In the case of help, it is not always easy at first sight to identify the ties which unite each member, as some of the connections have become opaque. That is why it seems interesting to look at the semantics of help from a diachronic perspective in order to determine how less central meanings emerged (cf. e.g. Geeraerts [1983], [1997]).

31The first meaning that will be addressed here is ‘provide/serve’, which is encoded by the form help [NP] to [NP].

3. Helpserve

32The meaning ‘serve’ (often for food), which was illustrated by sentence (18) above, is related to the prepositional construction found for the expression of movement (20-21), which is still in use today:

(20) Alef helped him to bed, and bade him a good night. (COCA, 2009, FICTION)

(21) Soldiers helped her across the tarmac, as she was still recovering from gunshot wounds and her three-week imprisonment in Iraq. (COCA, 2003, MAGAZINE)

33These spatial uses of help can easily be derived from the prototypical meaning of the verb. The only difference between the two is that the process following help is here encoded elliptically with a prepositional phrase, rather than with an infinitive clause.

34The first instances of the prepositional construction date back to 1380, according to the OED (s.v. help verb). The meaning of this construction was not restricted to purely spatial uses, such as ‘help to reach’ (cf. 20-21), but also included cases with a more metaphorical movement, toward an object of some sort, which could be paraphrased as ‘provide’. The purely spatial uses have persisted, but the other types of uses, in which the NP following the preposition does not indicate a location, are now obsolete, as the following examples suggest:

(22) The Protector [told] him that if he would help him to the Sight of Mr. Howe, [...] he would then give his judgement. (Edmund Calamy, A Continuation of the Account of the Ministers..., 1727)

(23) There were the two Persons of the Kingdom the King was most obliged to, since they helped him to the Victory of Bosworth [...]. (M. Rapin de Thoyras, History of England, 1728)

(24) He had teiz’d my Mother 500 times to help him to a Wife. (Old Bailey, 1725)

(25) I should help you to success. (Dickens, Nicholas Nickelby, 1839)

35These cases, in which help [NP] to [NP] means ‘provide’ rather than ‘help to reach’, have completely disappeared, except for the restricted meaning ‘provide food/serve’. According to the OED, the first occurrence of this sense is exhibited in the following sentence:

(26) Shall I help you to a piece of Veal? (Guy Miege, The Great French Dictionary, 1688)

36As the unnaturalness of examples (22) – (25) demonstrates, the help [NP] to [NP] structure meaning ‘provide’ has now been replaced by infinitive clauses. These sentences would sound more acceptable to our modern ears if they were re-written as follows:

(22’) The Protector told him that if he would help him to see/meet Mr. Howe, he would then give his judgement.

(23’) There were the two Persons of the Kingdom the King was most obliged to, since they helped him to win at Bosworth

(24’) He had teiz’d my Mother 500 times to help him to find a Wife.

(25’) I should help you to succeed.

37In a similar vein, it is fruitful to compare the following utterances, pronounced 180 years apart.

(27) The Prisoner said in his Defence, that he enquiring of a Baker’s Man for Business, he promised to help him to a Place. (Old Bailey, 1717)

(28) The RECORDER suggested that Mr. Wheatley might help him to find work. (Old Bailey, 1896)

38The gradual replacement of help [NP] to [NP] by help + infinitive clause is illustrated in table 1, which shows the percentage of these two structures in the Old Bailey for two distinct periods.

Table 1. Proportion of help [NP] to [NP] versus help + infinitive in the Old Bailey

Help [NP] to [NP]

(purely spatial cases excluded)

Help + infinitive


59,09 %


40,9 %



5,7 %


94,3 %


39Table 1 shows that the help [NP] to [NP] structure has decreased dramatically over the course of time. Between 1700 and 1740, nearly 60 % of cases of the verb help were followed by to NP. This percentage decreased to less than 6 between 1800 and 1840. This decrease is therefore dramatic and it is highly significant statistically (p << 0,0001). Many uses exhibiting the [NP] to [NP] structure involved the expression of processes (cf. sight, victory and success in the above cases). Infinitival structures being more clausal than noun phrases, they were seen as more appropriate to encode such processes.

40With the meaning ‘serve’, on the other hand, it is an entity which is provided. Consequently, this specialised use was not affected by the syntactic change that concerned the expression of processes. Therefore, during the nineteenth century, the general meaning ‘help to obtain / provide’ decreased, while the sense ‘serve’ persisted, as the following table suggests.

Table 2. The semantics of the help [NP] to [NP] structure in the Old Bailey: provide vs. serve

General help [NP] to [NP]
meaning ‘provide’

Specialised help [NP] to [NP food]
meaning ‘serve’


73 %

27 %


56.5 %

43.5 %

41The sense ‘to serve (food)’ was completely absent from the Old Bailey in the period between 1700 and 1750 but it gradually increased during the early 19th century, reaching around 40 % of the cases of help [NP] to [NP] in the period between 1821 and 1840. In the OED, the first mention of the corresponding noun, helping, referring to the action of serving food or to a portion of food, dates back to the early nineteenth century (s.v. helping, noun).

42On the whole, occurrences of helpserve involve reflexive pronouns, rather than simple pronouns. The following graph, based on COHA, shows the raw figures for the following string: help* [non-reflexive personal pronoun] to [singular noun / a / some]. Only the instances involving food and drink are mentioned here.

Figure 1. Occurrences of help(s/ing/ed) + [non-reflexive personal pronoun] + to + [singular noun / a / some] in COHA

Figure 1. Occurrences of help(s/ing/ed) + [non-reflexive personal pronoun] + to + [singular noun / a / some] in COHA

43As can be seen, the meaning ‘serve’ only marginally occurred with non-reflexive pronouns in COHA and this type of pronoun has now virtually disappeared.

44When the structure involves a reflexive pronoun, the verb predominantly encodes a transfer of food or drink. A study on the string ‘help* [reflexive pronoun] to’ in COCA shows that the percentage of food and drink is 75.8 % (385/508); the rest is composed of other types of supplies, notably money (4.3 %), vehicles (1.2 %), seats (1.2 %) and clothes (0.8 %).

45The second non-prototypical meaning of help that will be addressed is helpavoid.

4. Helpavoid

46An examination of an Early Modern English corpus, such as the Lexicons of Early Modern English (LEME) (Lancashire 1999), makes it possible to document the appearance of the meaning of helpavoid. This corpus gathers dictionaries and technical lexicons, as well as medical treatises, which proved key to understanding the semantic shift between prototypical help and helpavoid.

4.1. The emergence of helpavoid

47In Early Modern English, a monotransitive use of help started to appear alongside the help + NP + infinitive structure. From the original meaning of help derived the sense ‘facilitate’, present in (29), and ‘alleviate/cure’, present in (30):

(29) Myrtles are [...] verye profitable to helpe the restoring of broken bones. (Lanfranco of Milan, Chirurgia, 1565)

(30) If it be strened, it helpeth and swageth the sores in a mannes mouthe. (Anonymous, Banckes’s Herbal, 1525)

48This meaning is still in use, as indicated by example (16) above.
Several utterances included in LEME show how the sense ‘alleviate/cure’ could have led to a notion of protection or prevention. This is notably the case of (31):

(31) The floures of Lavander [...] mixed with Cinnamon, Nutmegs, and Cloves, made into pouder, and given to drinke in the distilled water thereof, doth helpe the panting and passion of the heart. (Thomas Johnson, The Herbal or General History of Plants, 1633)

49It is possible to imagine that such a treatment, used primarily to cure tachycardia and shortness of breath, was subsequently used to stave off the occurrence of such symptoms, as nothing in the utterance explicitly specifies the moment when the drug is to be used: at T-1 or at T+1 (T symbolising the moment when the dysphoric situation occurred). Similarly, in the following sentence, with a prepositional object, it is not made explicit whether the medicine is used for its curative or for its preventive properties.

(32) It helpeth against the bitings of any venomous beast, either taken in drinke, or outwardly applied. (Thomas Johnson, The Herbal or General History of Plants, 1633)

50Although this medicine is probably supposed to be used after a bite, it is not specifically encoded in the sentence. In addition, the use of the preposition against resembles the structure of the verb protect and this makes it possible for hearers/readers to infer that the medicine can be used preventively. One can therefore suggest that the lack of semantic explicitness of such structures, together with the frequent use of therapeutic drugs for prophylactic purposes, may have led to the appearance of helpavoid.

51Both prototypical help and helpavoid imply the existence of some dysphoric state of affair. The common semantic features between them can be described as an ‘action with a euphoric outcome’. Whether an unfortunate situation was solved after it happened (at T+1) or solved by avoiding its occurrence altogether (at T-1) was considered irrelevant and was conceptually backgrounded at the time when the semantic shift took place.

52This analysis confirms that the focus of speaker-oriented theories on the act of communication itself is well-founded. By studying utterances, rather than mere sentences, it becomes possible to uncover the decisive role that communication plays on the evolution of the lexicon. The interaction between speaker and co-speaker, and the various phenomena that it implies, such as potential semantic and structural ambiguity, as well as pragmatic inferences, are at the heart of language change (cf. Traugott & Dasher [2002]).

53Additionally, the preposition against, used in (35), substantiates the tenet that meaning emerges through a combination of lexical and grammatical items. This rather marginal prepositional use of help contributes to expanding the semantic interpretation of the verb by imperceptibly shifting its meaning towards that of other verbs used in similar prepositional constructions.

54Finally, the meaning of helpavoid became clearly distinct from ‘alleviate/cure’, as can be seen in the unambiguous sentence that follows:

(33) I do not believe the prisoner could help the accident. (Old Bailey, 1834)

4.2. The lexicalisation of can(-ed) not helpavoid

55Since the beginning of Late Modern English, helpavoid has been associated with the notions of capacity and negation, forming a rather loose colligation, i.e. a recurring combination of lexical items and grammatical features (Firth [1957]; Sinclair [1991]). Subsequently, the number of expressions used to refer to capacity gradually narrowed, as the following examples, which are no longer attested, show:

(34) This was a heavy piece of news to my nephew, but there was no way to help it but to comply. (William Defoe, Robinson Crusoe, 1719)

(35) [T]o say all in a word, he doats upon you; and I begin to see it is not in his power to help it. (Samuel Richardson, Pamela, 1740)

(36) It is impossible to help lamenting the unhappiness of Sir Edward. (Charlotte Smith, Ethelinde, 1789)

  • 3 Negative polarity items never occur in positive assertions using the simple past (Giannakidou [1998 (...)

56These examples demonstrate that non-prototypical expressions of inability have decreased over time. Additionally, the verb ceased to be a real negative polarity item (NPI). In the 19th century, the use of can help was licensed by a wide range of co-texts, like other typical NPIs3. For instance, it could be licensed by a subject containing a negative quantifier (37), by a wh-question (38) or by a than-clause (39):

(37) No man, outside of a deaf and dumb asylum, who is awakened at midnight on the 3d of July with – a hideous din that he knows will grow worse and worse for the next twenty-four hours, can help feeling that the Declaration of Independence was a terrible mistake, and that slavery and quiet are infinitely preferable to freedom and fire-crackers. (COHA, 1876, NEWS)

(38) It was difficult to understand how the snow could help melting in such an atmosphere. (COHA, 1832, NON FICTION

(39) I can no more help loving you than I can help breathing. (COHA 1885, FICTION)

57These types of contexts have dramatically decreased over time and the use of can help now tends to be limited to cases when it is in the syntactic scope of a negative particle (Pinson [2011]).

58As the means of expressing capacity became fixed and as the verb ceased to be attested in many NPI licensing co-texts, it lost most of its flexibility. All of these elements suggest that can (-ed) not help has undergone a process of lexicalisation: “the change whereby in certain linguistic contexts speakers use a syntactic construction or word formation as a new contentful form with formal and semantic properties that are not completely derivable or predictable from the constituents of the construction or the word formation pattern” (Brinton & Traugott [2005: 96]). Can (-ed) not help is now completely lexicalised, as it is virtually impossible to use never instead of not or to use intervening material in the middle of the multi-word expression (see Pinson [2011] for a more detailed analysis of this process).

59The lexicalisation of can(-ed) not helpavoid facilitates communication, as the use of unanalysable chunks improves fluency and gives the speaker more time for language planning (cf. Bolinger [1975]; Kuiper [1996]) and they are easier to interpret, since they can potentially limit cases of semantic ambiguity.

60In its lexicalised version, can(-ed) help means that somebody performs an action even though they would rather not. The euphoric component associated with the non-realisation of the action is still valid today. For example, a speaker may say I couldn’t help but laugh during a solemn celebration or I can’t help thinking that… , followed by a comment which is unpleasant for the co-speaker or for a third party.

61With this lexicalised multi-word verb, the subject is deagentivised and presented as an undergoer (patient or experiencer). This use is therefore far from being pragmatically neutral as it is meant to excuse the referent of the subject. The process being presented as unavoidable, it implies that the subject is not responsible for it (see Pinson [2011] on the pragmatic specificities of the can(-ed) help but V structure).

5. Helpcontribute

62The euphoric denotation expressed by the verb help is present in all the different types of help mentioned so far. However, other uses can be found today, as Mair [1995] noted:

(40) Nor have they eliminated the unburned hydrocarbons which help produce the smog that blankets such a motor-ridden conurbation as Los Angeles. (quoted in Mair [1995: 125])

(41) Negro cabbie John W. Smith, whose arrest for tailgating a police car [...] helped spark five days of rioting [...], was found guilty of assaulting a policeman. (quoted in Mair [1995: 125])

63These recent examples are particularly striking as they do not exhibit the same euphoric denotations as the other uses of help.

64In order to have a better understanding of this peculiar usage, I have looked for instances of the string ‘helped (to) create’ in COHA. I chose this particular string because it appears more frequently than ‘helped (to) spark’ and ‘helped (to) produce’, and because, like produce, the verb create can be followed both by negative and by positive referents.

5.1. Helped (to) create: a case study

65As Deléchelle [1998] noticed, structures with no NP after help are increasing. The following graph represents the raw figures for the string helped create and helped to create in COHA.

Figure 2. Occurrences of helped to create (in dark grey) and helped create (in light grey) in COHA

Figure 2. Occurrences of helped to create (in dark grey) and helped create (in light grey) in COHA

66The meaning ‘contribute’ correlates with a monotransitive structure, which can be associated to cases of prototypical help with a recipient omitted, as in the following example:

(42) Jear [rope], a piece of a hauset (in great ships) fastend to the main yard and foreyards, to help to hoyse up the yard and succor the ties. (Elisha Coles, An English Dictionary, 1676)

67The recipient can be omitted because it is inferable or generic, as in (42), where it refers to sailors.
The compatibility with dysphoric outcomes, however, is a more recent development, even though a few isolated cases of
help with a dysphoric outcome can be found in Early Modern English, in different types of structures:

(43) Assez escorche qui tient le pied: Prov. He does ill ynough that helps to doe euill. (1611, Randle Cotgrave, A Dictionary of the French and English Tongues)

(44) At last, I went to Sea of my own Accord, against my Mother’s Will; which I fear was the Cause of breaking her Heart, and has been a Means of helping me to this unfortunate End, because I disobeyed her. (Old Bailey, 1749)

68In spite of these early examples, it seems that the use of help followed by a dysphoric outcome remained extremely marginal before the 19th century. A look at COHA may be useful to understand how the compatibility with dysphoric results became possible.

69First, some examples show that it is not easy to classify utterances in terms of euphoricity or dysphoricity, because speakers may not agree on the axiological evaluation of the outcome:

(45) The impression is spreading here that the Soviet Union and the United States, far from having shaped alliances in their own images, have really helped to create an abundance of nationalisms that now threaten to engulf both in a display of old-fashioned European diplomacy played out over a worldwide scale. (COHA, 1963, NEWS)

(46) Thus, to say (in implied criticism) that the free alliances "have really helped to create an abundance of nationalisms" is in fact to state that the alliances have achieved success. (COHA, 1963, NON FICTION)

70Secondly, a substantial proportion of utterances with an animate subject contain outcomes which were not negative at first, but turned out to be dysphoric subsequently:

(47) The epidemic of strikes spreading throughout the country indicates that workers will no longer be the pawns of an acquisitive society and the victims of the industries they have helped to create. (COHA, 1937, NEWS)

71The creation of the industries was not necessarily a bad thing for workers at first, since it provided them with jobs and money, and maybe a feeling of achieving something, of being part of technological progress. It is only at a later stage that the industries started to be dysphoric for them.

72Another clear illustration of a chronologically ambivalent axiological evaluation of the outcome is provided by the following example, given here in context:

(48) Indeed, it was a part of Mr. Pulitzer’s instinct for polities and sagacity in reading the future in the present that he should have been so often on the right side of important public questions. It is a different matter when we consider the type of journalism which Mr. Pulitzer first fully exemplified in New York. […] He came to feel that what he had helped create had somehow become too powerful and had got away. […] From the first days of his control of the World, he struck out boldly for a new kind of “publicity” in journalism. But that word may mean many things. It covers not only the exposure of political scoundrels, but it denotes, in practice, a kind of constant clamor rising into a shriek, with a daily prying into private affairs as if nothing existed anywhere which should not be dragged into the light of day, and worse than all, a lurid exploitation of crime and nastiness. (COHA, 1911, MAGAZINE)

73As can be seen, Pulitzer was unable to foresee the future excesses of investigation journalism. The outcome of the process is therefore ambivalent, as the axiological evaluation varied across time, from positive to negative.

74The idea behind these examples is that the person responsible was not aware of the negative consequences that the action would subsequently have. The use of help thus highlights the fact that the helper unwittingly contributed to something which turned out to be dysphoric.

75This aspect is present in the following example, which underlines the fact that the dysphoric situation in question is happening unbeknownst to the public:

(49) For when we fail to practice tolerance in our individual lives, we jeopardize the structure for all. Insidiously, intolerance grows in its place. This does not happen suddenly, but by imperceptible degrees: a privilege is shorn away, a censorship erected, a hatred takes root, oppressive legislation is enacted. If we passively connive in this process, neglecting to protest unless our own rights and opinions are attacked, we shall discover some day that despotism has crept up on us – that we are about to be swallowed one and all by the monster we have helped create. (COHA, 1938, MAGAZINE)

76This point is also put forward in a few examples which evoke the concept of (tragic) irony. This literary notion refers to the fact that a character is unaware of an unfortunate state of affairs that the reader/spectator knows about:

(50) Ironically, some actions taken by the park service have helped to create problem bears. One example in recent years was the approval given by the park service for construction in Yellowstone of Grant Village, a mammoth complex that includes lodging units for nearly three thousand people, along with parking lots, recreational facilities, restaurants, a sewage treatment plant, a post office, and several other facilities. It was built on the confluence of several trout streams, a place that until the construction of Grant Village was a primary grizzly feeding site. Doubtless bears returning to the area in search of nourishment will come too close to people and will be dispatched as problem animals. (COHA, 1989, NON FICTION)

(51) Once humans decided to do anything but walk, once they became “traffic,” they had to learn a whole new way of getting around and getting along. What is the road for? Who is the road for? How will these streams of traffic flow together? Before the dust kicked up by the bicycles had even settled, the whole order was toppled again by the automobile, which was beginning to careen down those same “good road” the cyclists themselves, in a bit of tragic irony, had helped create. (COHA, 2008, NON FICTION)

77In view of these examples, one may hypothesise that it is through these ambivalent cases, in which the referent of the subject is unwittingly responsible for an unfortunate state of affairs, that the innovative meaning of help arose.

5.2. Grammaticalisation hypothesis

78According to Mair [1995]; [2002]; [2006], this semantic innovation is a sign that the verb help is undergoing a process of grammaticalisation, but it is actually difficult to prove or disprove this hypothesis. The semantic broadening and the increase in token frequency and type frequency described by Mair are real, but help still behaves like normal verbs in many respects, notably because it does not exhibit any of the NICE properties (cf. Deléchelle [1998]; McEnery & Xiao [2005]).

79Mair does not define the category that he calls “concatenation or modal idiom” [2006: 139] and no syntactic test is offered to identify this category. The only test mentioned is the replacement of the verb by a preposition, which is supposed to illustrate semantic broadening: “A ‘plan to help youngsters find jobs’ is nothing more specific than a ‘plan for them to find jobs’; and money used to ‘help pay’ for something is just money used for paying something” [1995: 265].

  • 4 G. Rohdenburg, “Aspecte einer vergleichenden Typologie des Englischen und Deutschen”, in C. Gnutzma (...)

80However, this test, from Rohdenburg [1990: 146-1484], seems problematic for two reasons. First, the fact that a word is acquiring the same distribution as a preposition certainly does not make it a modal auxiliary. Secondly, if this test was really valid, it would mean that the following verbs were also grammaticalising, which is doubtful:

(52) David McCreary is coming on to replace Tommy Jackson. [… coming on for / instead of Tommy Jackson]. (Rohdenburg 1990: 146-148)

(53) I’ve worked very hard to write this book.
(53’) I’ve worked very hard
on this book.

(54) He helped me (to) get over my addiction.
(54’) He helped me
with my addiction.

81Therefore, the only element that is decisive in establishing that help does not behave syntactically like prototypical lexical verbs is the omission of to.

82When a lexical item becomes grammaticalised, one usually observes a phenomenon known as layering (Hopper [1991: 22]). This means that the grammaticalised item and the same item in its ungrammaticalised version can coexist. For example:

(55) I’m going to London next week.

in which going to isn’t grammaticalised, coexists with

(56) John’s not going to be happy about it.

in which it is grammaticalised.

83When grammaticalisation is accompanied with phonetic erosion, it does not concern the non-grammaticalised item (Bybee [2002]):

(55’) *I’m gonna London next week.

84Yet, in the case of help, the omission of to initially concerned cases with animate subjects, in which the verb really has its prototypical meaning, as in:

(57) The Prosecutor’s Wife depos’d, That the Prisoner being employed to help her Wash, carried away the Goods. (Old Bailey, 1717)

(58) Mrs. Taylor her Mistress deposed, that on Sunday the 26th of October, the Prisoner complain’d that she was not well, and had the Gripes, &c. but did help her make the Bed, and she afterwards went out to Church leaving her to dress Dinner. (Old Bailey, 1718)

(59) He did not confess to me any more Robberies, but own’d a great number of small Thefts, he constantly did as occasion offer’d; and in Alleviation of his Guilt in this Affair, he said, he was no farther concern’d in the Robbery, only that he was present at the Time of committing it, he being 7 Yards distant, and did not help them take any Thing from Mr. Daniel. (Old Bailey, 1742)

85The high compatibility between an animate subject, who actually brings help in the literal sense, and the omission of to has been noticed by many researchers, as mentioned above. So, how could the omission of to be a sign that help is grammaticalising if it originally affected only the prototypical meaning of help?

86Finally, Mair considers that “adding to before the infinitives would be slightly incongruous” in cases which display negative effects [2002: 126]. Yet, the expression of a negative outcome does not necessarily coincide with the omission of to. In my study on helped (to) create, I even notice a higher proportion of negative outcomes in combination with the use of to than with the omission of to.

Table 3. Proportion of negative outcomes following helped create and helped to create in COHA

Proportion of negative outcomes

“helped create”

26 %

“helped to create

42,7 %

87Therefore, Mair’s hypothesis of grammaticalisation, however tempting, is slightly undermined by these observations on the omission or presence of to.

88After this analysis of the emergence of the non-prototypical meanings of help, it could be useful to provide a grid which synthetises the similarities and the differences between the various meanings of the verb.

6. Synthesis

89The following table is meant to capture the diversity of help by means of a restricted set of features.

Table 4. A description of the polysemy of help













S + V + {recipient} + {themeprocess}
e.g. He helped me (to) do it.

solved at T+1





+ specific: lack

Si + V + {recipienti} + {theme}
e.g. He helped himself to some tea.

transfer at T+1






S +
can-(-ed) not V + {themeprocess}
e.g. We couldn’t help laughing .

solved at T-1






S + V + {theme}
e.g. The drug helped the pain.

solved at T+1






S + V + {theme=process}
e.g. They helped (to) produce a smog.

90Table 4 provides a semantic description of the various meanings of help based on a selection of distinctive features and it displays the correspondence between these meanings and their respective argument structure. Although all the senses of help involve an action of some kind related to a situation x, not all of them express an action resulting in a new situation, since helpavoid implies the absence of change. Similarly, the meaning ‘contribute’ differs from all the others as it does not involve the resolution of some problematic situation.

91The semantic evolution of the verb could be described as a movement from ‘interaction’ to ‘action’, with a combination of parameters such as [+/- dysphoric situation] and [dysphoric situation solved at T+1 or at T-1]. In the most recent meaning of help, the notion of interaction has diluted into a mere denotation of shared action, i.e. the origin of the process is one cause among other potential factors. This is virtually the only trace of the prototypical meaning of help left, as the euphoricity component is no longer applicable here.

92These observations can be summarised as follows, in a graph representing the semantic evolution of the verb help.

Figure 3. The semantic evolution of the verb help

Figure 3. The semantic evolution of the verb help


93The polysemous verb help displays a wide variety of meanings, some of which are even contradictory. It is possible to find some semantic core that encapsulates all of them, because they all refer to an action related to a given situation. However, this scenario is so general that it fails to really account for what the verb help actually means. I have therefore adopted a diachronic perspective, as it seems better suited to explain what each secondary meaning has in common with the prototypical one.

94Quite clearly, syntax and semantics are interrelated, as each structure reflects a different meaning. The first monotransitive use, meaning ‘alleviate’, which still exists today, gave rise to the sense ‘avoid’, by pragmatic inference: ‘if x cures y, then x can also be used to avoid y’. The second monotransitive use gave rise to the neutral meaning ‘contribute’, which is compatible with dysphoric outcomes. The elliptical prepositional version of prototypical help, with the rather broad meaning ‘assist in reaching or obtaining’, split into two uses, the spatial use and the meaning ‘serve’.

95The fact that meanings are so divergent can partly be explained by the distinct phenomena which have been at play. Semantically speaking, we observe a case of narrowing (help to obtain => serve) and a case of broadening (help => contribute). Syntactically speaking, we note that helpavoid underwent lexicalisation, while helpcontribute may correspond to a case of grammaticalisation, although this hypothesis remains hard to prove.

96The strict correspondence between form and function probably explains why it is possible to have such a wide array of meanings for the same word. No two uses share the same argument structure and the same type of complementation. Thus, maximally isomorphic constructions compensate for the lack of isomorphism of the verb.

Top of page


Adamczewski Henri & Delmas Claude, 1993 [1982], Grammaire linguistique de l’anglais (4e édition), Paris: Armand Colin.

Benveniste Emile, 1966, Problèmes de linguistique générale, Paris : Gallimard.

Bolinger Dwight, 1975, Aspects of Language, New York: Harcourt, Brace, Jovanovich.

Bolinger Dwight, 1979, “The Jingle Theory of Double -ing”, in Allerton D. J., Carney Edward & Holdcroft David (eds). Function and Context in Linguistic Analysis: A Festschrift for William Haas, Cambridge: Cambridge University Press, 41-56.

Brinton Laurel J. & Traugott Elizabeth Closs, 2005, Lexicalization and Language Change, Cambridge: Cambridge University Press.

Bybee Joan, “Cognitive processes in grammaticalization”, 2002, in Thomasello Michael (ed.), The New Phychology of Language volume II. New Jersey: Lawrence Erlbaum Associates Inc.

Cadiot Pierre, 1994, « Représentations d’objets et sémantique lexicale : qu’est-ce qu’une boîte ? ». French Language Studies 4, 1-23.

Chomsky Noam, 1957, Syntactic Structures, The Hague/Paris: Mouton.

Chomsky Noam, 1980, Rules and Representations, New York: Columbia University Press.

Chuquet Jean, 1986, TO et l’infinitif anglais, Cahier de recherche en grammaire anglaise n°3, Paris : Ophrys.

Cotte Pierre, 1982, « To, opérateur de dévirtualisation », Modèles linguistiques 4.2, 135-49.

Culioli Antoine, 1990-1999, Pour une linguistique de l’énonciation, Paris/Gap : Ophrys.

Davies Mark, The Corpus of Contemporary American English (COCA): 410+ million words, 1990-present, 2008-, Available online at

Davies Mark, The Corpus of Historical American English (COHA): 400+million words, 1810–2009. 2010-, Available online at

Delechelle Gérard, 1998, « Help : syntaxe et sémantique », Cycnos, Actes de l’atelier de linguistique du congrès SAES de 1997 vol. 15,117-129.

De Vogue Sarah, 2004, « Syntaxe, référence et identité du verbe filer », LINX 50, 135-167.

Dixon Robert M. W., 1991, A New Approach to English Grammar on Semantic Principles, Oxford: Clarendon Press.

Duffley Patrick, 1992, The English Infinitive, London: Longman.

Firth John Rupert, 1957, Papers in Linguistics 1934-1951, London: Oxford University Press.

Fischer Olga, 1995, “The distinction between TO- and bare infinitival complements in Late Middle English”, Diachronica 12, 1-30.

Franckel Jean-Jacques & Lebaud Daniel, 1995, « Les échappées du verbe sentir », Langues et Langage. Problèmes et raisonnement en linguistique. Mélanges offerts à Antoine Culioli, dir. Bouscaren Jeanine, Franckel Jean-Jacques & Robert Stéphane, Paris : PUF, 261-77.

Franckel Jean-Jacques & Paillard Denis, 2007, Grammaire des prépositions, Paris : Ophrys.

Geeraerts Dirk, 1983, “Prototype theory and diachronic semantics. A case study”, Indogermanische Forschungen 88, 1-32.

Geeraerts Dirk, 1997, Diachronic Prototype Semantics: A Contribution to Historical Lexicology, Oxford: Clarendon Press.

Giannakidou Anastasia, 1998, Polarity Sensitivity as (Non)veridical Dependency, Amsterdam/Philadelphia: John Benjamins.

Girard Geneviève, 1996, « Faut-il faire l’hypothèse d’un invariant différentiel ? ». Modèles Linguistiques 17,119-132.

Haiman John, 1994, “Ritualization and the development of language, in Pagliuca William (ed.), Perspectives on Grammaticalization, Amsterdam: John Benjamins, 3-28.

Haspelmath Martin, 1989, “From purposive to infinitive – a universal path of grammaticization”, Folia Linguistica Historica 10, 287-310.

Hitchcock Tim, Shoemaker Robert, Emsley Clive, Howard Sharon & McLaughlin Jamie et al. The Old Bailey Proceedings Online, 1674-1913 (, version 7.0, 24 March 2012).

Hopper Paul J., 1991, “On some principles of grammaticalization”, in Elizabeth Traugott Closs & Heine Bernd (eds), Approaches to Grammaticalization vol. 1. Amsterdam/Philadelphia: Benjamins, 17-35.

Jackendoff Ray, 1990, Semantic Structures, Cambridge, Mass.: MIT PRESS.

Joly André, 1980, « But : morphème de la subordination de l’anglais », Travaux de linguistique et de littérature XVIII. 1,269-285.

Kerbrat-Orecchioni Catherine, 2002 [1999], L’Enonciation 4e ed. Paris : Armand Colin, 2.

Kleiber Georges, 1999, Problème de sémantique, la polysémie en question, Villeneuve d’Asq : Presses Universitaires du Septentrion.

Kuiper Koenraad, 1996, Smooth Talkers: The Linguistic Performance of Auctioneers and Sportscasters. Mahwah, NJ: Erlbaum.

Lakoff George, 1987, Women, Fire, and Dangerous Things: What Categories Reveal about the Mind, Chicago: University of Chicago Press.

Lancashire Ian, 1999, Lexicons of Early Modern English, University of Toronto Press,

Los Bettelou, 2005, The Rise of the To-Infinitive, Oxford/New York: Oxford University Press.

Maingueneau Dominique, 1976, Initiation aux méthodes de l’analyse du discours, Paris : Hachette.

Mair Christian, 1990, Infinitival Complement Clauses in English: A Study of Syntax in Discourse, Cambridge/New York: Cambridge University Press.

Mair Christian, 1995, “Changing patterns of complementation, and concomitant grammaticalisation, of the verb help in present-day British English”, in Aarts Bas & Meyer Charles F., The Verb in Contemporary English: Theory and Description, Cambridge: Cambridge University Press, 258-272.

Mair Christian, 2002, “Three changing patterns of verb complementation in Late Modern English”, English Language and Linguistics 6.1, 105-31.

Mair Christian, 2006, Twentieth-Century English: History, Variation, and Standardization. Cambridge/New York: Cambridge University Press.

McEnery Tony & Xiao Richard, 2005, “Help or Help to: what do corpora have to say?”, English Studies 86.2, 161-87.

Paillard Denis, 2001, « A propos des verbes “polysémiques” : identité sémantique et principes de variation », Syntaxe et Sémantique 2, 99-120.

Pauly Emilie, 2010, La Polysémie – Réflexion théorique, méthodologique et application à la lexicographie : l’exemple des verbes aller, partir et tirer en français contemporain, Paris : L’Harmattan.

Picoche Jacqueline, 1995, Structures sémantiques du lexique français, Paris : Nathan.

Pinson Mathilde, 2009, Norme et variation en anglais contemporain – Etude variationniste de quelques subordonnants: complémentation de help et like conjonction. Thèse de doctorat. Université Paris 3.

Pinson Mathilde, 2011, “The paradoxical success of can (-ed) not help but V: when the extragrammatical meets the non-compositional”, in Koenraad Kuiper (ed.). Yearbook of Phraseology volume 2, 17-44.

Puckica Jérôme, 2008, « Sens et relations de sens dans les grammaires de construction ». Lexis, E-Journal In English Lexicology 1 : La polysémie/Polysemy.

Quirk Randolph, Greenbaum Sidney, Leech Geoffrey & Svartvik Jan, 1985, A Comprehensive Grammar of the English Language, London/New York: Longman.

Rohdenburg Günter, 2003, “Cognitive complexity and horror aequi as factors determining the use of interrogative clause linkers in English”, in Rohdenburg Günter & Mondorf Britta, Determinants of Grammatical Variation in English. Berlin/New York: Mouton de Gruyter, 205-249.

Rosch Eleanor, 1973, “Natural categories”, Cognitive Psychology 4 (3), 328-350.

Schlüter Julia, 2003, “Phonological determinants of grammatical variation in English: Chomsky’s worst possible case”, in Rohdenburg Günter & Mondorf Britta, Determinants of Grammatical Variation in English. Berlin/New York: Mouton de Gruyter, 69-118.

Simpson John A. & Weiner Edmund S. C. (eds.), The Oxford English Dictionary on CD-ROM, 2nd edition, Oxford: Oxford University Press.

Sinclair John, 1991, Corpus, Concordance, Collocation, Cambridge: Cambridge University Press.

De Swart Henriette, 1998, “Licensing of negative polarity items under inverse scope”. Lingua 105(3–4), 175–200.

Talmy Leonard, 2000, Toward A Cognitive Semantics, vol. 1, Conceptual Structuring Systems. Cambridge, Mass./London: MIT Press.

Toupin Fabienne & Lowry Brian, 2010, « L’Invariant à l’épreuve de la diachronie », Coréla 8.2,

Traugott Elizabeth Closs & Dasher Richard B., 2002, Regularity in Semantic Change, Cambridge: Cambridge University Press.

Victorri Bernard & Fuchs Catherine, 1996, La Polysémie : Construction dynamique du sens, Paris : Hermès.

Wierzbicka Anna, 1988, The Semantics of Grammar, Amsterdam / Philadelphia: John Benjamins.

Top of page


1 It is difficult to be very specific regarding the date when this change occurred. Indeed, it is still possible to use purpose clauses with help today and opinions still vary concerning the status of infinitive clauses after this verb (see Mair [1990] and Quirk et al. [1985]). However, the diachronic connection between purpose clauses and infinitive complements is well documented (e.g. Los [2005] and Haspelmath [1989]). The following sentence shows that this change had been established by at least the late sixteenth century in the case of help: What Letters did he help the Frenche Ambassador to conveye into Scotlande? (Collection of State Papers relating to Affaires in the Reign of Queen Elizabeth [1571], published in 1759 by Cecill William).

2 The near-synonyms given here are used to illustrate the different meanings of help. They are not necessarily commutable with help in context because help displays several syntactic peculiarities: e.g. the omission of to does not occur with other verbs of assistance and the use of the complementiser but has completely disappeared in unmarked contemporary English, except after cannot help (see Pinson [2011] on the use of but after cannot help).

3 Negative polarity items never occur in positive assertions using the simple past (Giannakidou [1998]). By contrast, such items occur when they are in the syntactic scope of a) a negative particle, b) a restrictive particle, c) a subject containing a negative quantifier, d) a subject containing a restrictive quantifier. They are also licensed by e) a negation with inverse scope i.e. when the item is “within the semantic scope of an expression which does not c-command it at S-structure” (de Swart [1998: 179]), f) wh- and yes/no questions, g) if-clauses, h) than-clauses and i) before-clauses.

4 G. Rohdenburg, “Aspecte einer vergleichenden Typologie des Englischen und Deutschen”, in C. Gnutzmann, Kontrastive Linguistik, Lang, Frankfurt, 1990, 133-152.

Top of page

List of illustrations

Title Figure 1. Occurrences of help(s/ing/ed) + [non-reflexive personal pronoun] + to + [singular noun / a / some] in COHA
File image/png, 23k
Title Figure 2. Occurrences of helped to create (in dark grey) and helped create (in light grey) in COHA
File image/png, 32k
Title Figure 3. The semantic evolution of the verb help
File image/png, 16k
Top of page


Electronic reference

Mathilde Pinson, « On the polysemy of help », Lexis [Online], 9 | 2015, Online since 13 May 2015, connection on 25 March 2017. URL : ; DOI : 10.4000/lexis.943

Top of page

About the author

Mathilde Pinson

Université Sorbonne Nouvelle – Paris 3, SeSyLiA

Top of page
  • Logo Centre d’études linguistiques
  • Logo Université Jean Moulin – Lyon 3