Skip to navigation – Site map

Calamities and Counterfactuals: A Historical View of Polarity Reversal

Debra Ziegeler


Previous studies have attributed little attention to the historical factors surrounding the development of counterfactual meanings in the adverbs almost and nearly, though some have referred to evidence of pleonastic negation found in proximatives across languages. In the present study, the historical development of the two adverbs will be considered in investigating recent data from Late Modern English, in which a majority of counterfactual uses appear with a complement referring to undesirable events, sometimes expressed hyperbolically. It is hypothesised that the presence of intersubjectivity contributed significantly to the development of counterfactual meanings in both adverbs, in focusing attention on the aversion of, rather than the proximity to, the event described in the complement. However, in the case of nearly, the proximative meanings of counterfactuality are likely to have emerged out of contexts in which meanings of temporal proximity (expressing counterfactuality) were ambiguous with meanings of spatial proximity (expressing factuality), as in the frequent appearance of nearly with past participles.

Top of page

Author's notes

Part of the present study was presented at the 18th International Conference on English Historical Linguistics, July 14-17, 2014, KU Leuven. Acknowledgements are due to the audience at that conference and to EA 4398, Université Paris 3, for sponsoring the earlier presentation of the paper at the University of Perpignan, October 2013. Acknowledgements are also due to Olivier Simonin and the members of the audience at Perpignan for many interesting comments and suggestions, and to Geneviève Gillet for translating the abstract.

Full text


1It is an almost forgone conclusion in semantic circles today that the semantics of almost and similar proximative devices in other languages refer to an implicit negation which is present as highly conventionalised or even, according to some accounts, an entailed meaning of such functions. Discussion and debate on the precise semantics of the adverb almost in English have raged for over 30 years now (see, for example, the work of Sevi 1998, Hitzeman 1992, Horn 2002, 2011, Pons Bordería and Schwenter 2005, and Amaral 2007, who have claimed for an entailment analysis of the adverb, or Sadock 1981, Atlas 1984, 2005, and Ziegeler 2000, 2010, who argue for an analysis based on conversational or conventional implicature). Further research has revealed a large number of languages in which the meanings of negation are no longer implicit in the context, but are expressed overtly as a form of pleonastic negation (e.g. Wierzbicka 1986, Kuteva 1998, Pons Bordería and Schwenter 2005, and Amaral 2007). Schwenter 2002 and Pons Bordería and Schwenter 2005 have gone further to suggest that the presence of such expletive negation is evidence for an entailment analysis of almost, following Horn's (2002) analysis of the adverb as an unasserted entailment.

  • 1 The data and examples shown in sections 3.1 and 3.2 also appear in that study.

2It is not within the scope of the present paper to attempt to carry such arguments further, but it should be noted that little research has been undertaken, to present knowledge, on the diachronic development of the adverb, a possible case of 'adverbialisation' from what was formerly a complex quantifier meaning roughly 'mostly all'. The meanings familiar to its present-day use can in many cases be described as modal, since in collocation with preterite or past participle verb forms, they express a degree of counterfactuality, in that they are predictions of alternative outcomes relating to past events. The gradual emergence of such meanings has been discussed in Ziegeler (2015, to appear), and will be reviewed briefly in the present paper. 1 In nearly all such cases, the review of Late Modern English data reveals that the diachronic development of counterfactual meanings emerged largely due to the co-occurrence of almost with predicates referring to undesirable events and situations: in other words, it was not counterfactuality that was being expressed so much as counter-to-desire. In some cases, such predicates referred hyperbolically to extremely exaggerated circumstances.

3The present study opens the case even further, to study the parallel transformation of the adverb nearly into a counterfactual adverb, from what was, apparently, simply a manner adverb expressing closeness (the equivalent of 'closely' in today's English). The first section will briefly review earlier work on nearly; the second section will review the literature to date discussing the semantics of proximity of almost and comparing the two adverbs. The third section will present a large portion of data obtained from the Corpus of Late Modern English Texts (Extended Version) (CLMET(EV)), and covering the period 1710-1925. The data will be discussed in section 4, and a hypothesis presented, in order to explain the extraordinary orientation-switch in the polarity of the adverb nearly (when co-occurring with preterite verbs and past participles), from often entailing P (its predicate proposition) to reversing the entailment altogether to indicate counterfactuality or the non-occurrence of the event referred to in the predicate. The hypothesis will propose a multiple-orientation polysemous framework for the semantics of such proximatives.

1. Previous research on nearly

4As noted in the introduction, the literature on proximatives has been overwhelmed by studies focusing on the adverb almost, and often also barely (e.g. Amarel 2007, Horn 2002, 2011, and Ziegeler 2010), while nearly has been largely neglected in the current research. Horn's only (2011: 6) mentions of nearly refer to previous discussions in Sadock (2007) and van Dongen (1921), and he notes that almost and nearly are definitely not entirely synonymous, quoting Sadock (2007) as suggesting that nearly involves that some kind of expectation on the part of the speaker was exceeded, while almost is less likely to indicate this. The presence of such nuances may explain why the example cited from van Dongen (1921): Almost/#Nearly nobody was there does not sound as felicitous as it should: it suggests the speaker is expecting nobody to be there.

5The argument is raised again in Liberman (2007), prior to and cited in Sadock (2007), when, on the basis of some informal statistics gathered from Google sites, he finds that almost no-one is 3,840% more frequent than nearly no-one, and almost everyone is 30% more frequent than nearly everyone. However interesting these statistics may sound, their investigation requires a fuller consideration of the total frequency of the two adverbs across a corpus, and it may simply be the case that almost is more frequent in any context, negative or otherwise, thus reducing such statistics to a mere representational proportion of a general pattern. On the same website, Allan Wechsler is cited as noting the difficulty in quantifying nearly zero as being of a similar nature to the nearly no-one examples: it is usually considered that nearly approximates a degree or quantity slightly below the target referred to, and if the numerical value is zero or a negative value, it suggests that the value is below zero, somewhat more difficult to conjecture; this is also agreed upon by Sadock (2007). Almost, then, may differ from nearly in qualifying more generalised meanings of proximity (and so almost no-one can imply proximity above the target instead of below: i.e., one or two people).

6However, it would not be difficult to find weather reports forecasting temperatures going down to nearly zero in the winter, though, admittedly, the direction of the trajectory might also be indicated in such expressions. But not always. On an earlier page of the same website (2004), Liberman states:

(1) ... such tests will pick out far too many words and phrases whose expected frequency over the span of text in question is nearly zero.

  • 2 Consider also whose expected frequency over the span of text in question is almost zero, in which t (...)

7While linguists are renowned for their innovative use of language (see, e.g, inanimate whose in (1)), the use of nearly zero in (1) to quantify the frequency of words does not appear odd or infelicitous, since it is understood that the quantification of words or phrases refers only to those in existence, not those which are not yet existing.2 A random collocation search of nearly zero temperatures on Google yields 44,300,000 hits (21/6/14), the most recent referring to the polar summer (but nevertheless most likely a rise in temperatures rather than a fall). Nearly no + NP, though, is less frequently accepted, and Liberman (2007) finds only one example in a wide survey of historical English literature. Thus, according to Sadock (2007) and his colleagues, the differences between nearly and almost relate to the orientation tendencies of the two proximatives - the use of nearly conveying an implicit sense of expectation, while almost conveys no similar nuances at all.

8According to Sadock (2007), though, both almost and nearly qualify for negative entailment status under Horn's analyses, something with which Horn disagrees. Horn suggests that nearly n exceeds expectations or what was hoped for, while almost n does not conventionally connote any particular desire, hope or expectation, although supporting a conversational implicature to the same effect. The differences may be due to the different origins and developmental paths: nearly, to begin with, is a younger adverb than almost which derives from a quantifier plus determiner univerbation originating in Old English. Even though adverbs are doubtful cases for grammaticalisation analyses (since they are ambivalent as grammatical morphemes - see Brinton and Traugott 2005: 132), the developmental path followed by almost seems clearly to reveal an increasing tendency for more and more abstract uses to appear, unrelated to spatial or quantificational proximity. Nearly, on the other hand, was always an adverb, like barely, and its development as a proximative can be traced no further than Early Modern English times. The comparative historical development of the two adverbs will be reviewed below; meanwhile, it is necessary to consider the semantic range of meaning available to proximatives generally.

2. Semantic analyses of almost

2.1 Conjunctive analyses

9The present study does not permit sufficient scope to provide an overview of the entire, on-going debate between advocates of the entailment analysis of almost and those of the implicature account; the reader is referred to Horn (2011) for a complete summary. The entailment account of the semantics of almost, attributed above to Horn (2002, 2011), as well as Amaral (2007), Hitzeman (1992), Pons Bordería and Schwenter (2005), and Schwenter (2002), gave rise to the suggestion that the present-day meaning of almost was derived from two independent senses: the proximative sense and the polar sense (i.e., expressing negation). Such distinctions came to be termed the Conjunctive Analysis (e.g. Sevi 1998, Horn 2002, 2011, Amaral 2007, Pons Borderia and Schwenter 2005). It was therefore admitted that the sense of almost was subject to a division between its 'basic' meaning of proximity, and a derived meaning of negation, understood as its polar meaning, but not necessarily that the two meaning components were interrelated. The division of semantic labour was thus illustrated in Horn (2002: 62) in the following way, distinguished by whether the component was entailed or asserted, or both:

  • 3 Horn (2002) adapts this scale from Schwenter (2002), but both studies, being primarily concerned wi (...)





Proximal component



Polar component



10The polar component of negation was taken to represent an assertion that, by contrast to the proximative component, was described as inert, or 'silent' in the context (unasserted). It is almost incontrovertible that the polar meanings of negation are not asserted: almost (and any other proximative, for that matter) do not function as negators in the grammatical system of English. Furthermore, the polar meanings are not always so clearly visible in the case of almost, for example in (3), from the ICE-GB (w2a-028), discussed in Ziegeler (2010), where:


a. He was almost underwater


b. He was not underwater.

11In such examples, the entailment of negation is disputable, as many would disagree that (3a) entails (3b). The most curious of addressees might be obliged to ask how it is known that he was almost underwater, if he was not (even slightly) underwater, and illustrating clearly the element of prediction carried by proximatives and their polar meanings and their frequent use in quantifying partial representations of entire states or qualities. In examples such as (3), in which almost refers to a situation falling short of an assumed target, or absolute measure, an additional implication emerges wherein the spatial proximity to an assumed target becomes reinterpreted as a temporal proximity, and the target becomes predictable. It becomes questionable, then, whether such predictions can be classed as entailments.

  • 4 However, this type of scalar analysis may not necessary apply to all verbal predicates (for this re (...)

12In (3), then, the prediction of total submersion carries with it a certain sense of temporal directionality or orientation as well (see Ziegeler 2010). It also carries a certain assumption that the subject is at least partially underwater, and raising the question whether the polar component of meaning (i.e. suggesting the subject in (3) was not underwater) is strong enough to be labelled an entailment (the subject could not, in fact, be completely out of the water.) What (3a) does deny, however, is a counterpart expression using an adverb of completion (e.g., he was completely/quite/entirely underwater). Hence, (3a) implicates in a scalar fashion, he was not quite underwater, and he was not quite underwater implicates the assertion of its alternate, (3a). In this way, almost is seen as a focus adverb on a quantity scale such as <quite/entirely, almost>, where the negation of an informationally stronger item (e.g. quite/entirely) implicates the assertion of the other, informationally weaker item (almost).4

13 For present purposes, the Conjunctive Analysis may therefore be revised, referring in preference to a derivative analysis rather than a conjunctive one. This would then presuppose a semantic connection between the unrelated senses of proximity and polarity, and introduce an additional component of goal orientation to the description:





Proximal component



Polar component



Orientational component (prospective)



14From (4) it can be seen that not just one, but two inferred meanings may arise out of one meaning, which is both asserted and entailed. It should also be noted that the orientational component is prospective, indicating direction towards P, while the polar component of negation is indicative of contrary orientation: away from P. The inclusion of two contradictory elements of meaning is thus problematic for a conjunctive analysis, and requires a complete pragmatic explanation unrelated to the restrictions carried by entailments. It was proposed in Ziegeler (2010) that the inferences for the derived meanings of polarity and orientation were not entailments, but rather pragmatic implicatures emerging from the gradually expanding functions of almost over time.

2.2 The pragmatics of orientation in almost

  • 5 In fact, Nouwen (2006) supplies adequate information to suggest contradictions between temporal ori (...)

15Logically speaking, there is nothing in the meaning of proximity to P to entail that whatever is proximative to P is also necessarily moving in the direction of P.5 This is an additional meaning derived from the use of proximal adverbs with verbal predicates, as will be discussed below. It also means that the Conjunctive Analysis should be revised to account for two polar meanings: positive and negative, with a basic assumption that the only entailed and asserted component of meaning is ‘closeness’ to P. The two opposite orientational meanings are evidence of an orientation shift, or polarity shift, in the development of the adverb. Such a shift will be discussed at length below, and also in regard to nearly, a more recent addition to the family of proximatives in English (as noted above).

16Orientational opposition can also be seen with other polarised proximative adverbs: Horn (2008) also contrasts almost and barely, the latter understood as the 'reverse' of almost:

(5) The tank is almost half full – so {let’s drive on/#we’d better stop for gas}.

(6) The tank is barely half full – so {we’d better stop for gas/#let’s drive on}.

17The pragmatic orientation towards P in (5) and towards ~P in (6) is described as rhetorical by Horn; however, the connecting clauses suggest that the meaning is affected by nuances of desirability as well. In (5) the directional inferences in almost are conventionalised to the point at which the amount of petrol referred to may even be construed (albeit implausibly) as increasing; hence the optimism of the speakers. In (6) the orientation is towards the emptying of the tank, a more realistic, but less desirable situation though in (6) the tank is assumed to contain more petrol than in (5) (since barely P entails P – see, e.g. Sevi 1998).

18The orientational meanings are hypothesised to be derived via the Quantity maxims of Grice (1975), Quantity 2, renovated in the form of R-based implicatures in Horn (1989/2001) and Levinson's (2000) I-heuristic. Levinson's version of the original Gricean sub-maxim of Quantity extended the definition to include inferences to the stereotype, though it is doubtful that the definition most useful to the present study could also include such extensions. Nevertheless, the maxim requires that the hearer or addressee read more into the expression than what is actually uttered, and so the maxim calls for an enrichment at the informational level. Thus, the example in (5) clearly illustrates the type of informational enrichment associated with the operation of such maxims and principles: the orientational inference associated with the use of the proximative is that the tank is almost half-full and therefore tending towards a positive polarity reading rather than a negative one.

19Similar contrasting examples appearing in Horn (2002: 57) show a difference of time reference:

(7a) My printer is almost functional!

(7b) My printer was almost functional!

20Horn describes (7a) as 'good news' and (7b) as 'bad news', without referring to the possible interplay of Gricean and Neogricean Quantity implicatures ((7a) suggests 'it will be functional' (via Horn's R-based inferences and Levinson's I-heuristic). (7b), on the other hand, suggests it is not functional now, and thus exploits the Quantity 1 maxim of supplying as much information as is necessary (also revised as Horn's (1989) Q-based inferences and Levinson's (1995) Q-principle), according to which the maximisation of information supplied implicates the negation of anything left unsaid. The higher informational value of past time reference is due to the fact that the past is generally known, and so anything said about the past presumes that that was the maximum knowledge possessed by the speaker. The same Q-based inferences are available from the use of nearly appearing in a context with past time reference:

(7c) c. My printer was nearly functional!

21Like (7b), (7c) also expresses ‘bad news’ suggesting the printer no longer functions. The contrast in time reference is thus particularly relevant to the interpretation of counterfactuality in both the proximative adverbs, nearly as well as almost .

2.3 Reinforcement with negation

22It will be seen in the empirical investigations to follow that the orientation of counterfactual situations is related to an increasing tendency to use the proximative adverbs in expressing the avoidance of undesirable circumstances, something that was already observed in the so-called 'pleonastic' negation of proximatives in other languages, such as Chinese (Zhu 1956, Li 1976, Peyraube 1979, Biq 1989,) Polish (Wierzbicka 1986), Bulgarian (Kuteva 1998, 2001), and Spanish (Schwenter 2002, Pons Bordería and Schwenter 2005), e.g. (Pons Bordería and Schwenter 2005):

(8) Por poco se mata
‘She was almost killed’

(9) Por poco no se mata
‘She was almost/almost not killed’

23It is seen in (9) that the same expression, which includes a negative element, can derive two ambiguous meanings: one positive and one negative. The addition of the negative element was discussed in Ziegeler (2010) as a reinforcement of the inferences of negation produced by the proximative adverb alone in examples like (8). However, the addition of a negative element also carries with it the possibility of the speaker’s concern for the evaluations of the addressee, and the need to include reinforcing negation in the proximative expression has been attributed in Ziegeler (2015, to appear) to the operation of an intersubjective element of meaning in the use of proximatives; i.e., a desire to strengthen the expression of avoidance of a predicted, unpleasant past event.

24According to Traugott (2010), intersubjectivity was thought to refer to the speaker's linguistic accommodation of the addressee's self-image, 'face', attitudes and beliefs, and the definition could be expanded to suggest that the addressee's emotive 'face' is also important to the locutionary agent's use of language. This intersubjective element can be associated with earlier descriptions by Boucher and Osgood (1969) and Horn (1989/2001) of a phenomenon known as the 'Pollyanna Hypothesis', by which speakers tend to use positively-connoting meanings in preference to negatively-connoting expressions. Examples include the tendency, discussed in Horn (1989: 160) and Boucher and Osgood (1969: 4), for positive adjectives to be negated rather than negative ones (e.g. unhappy vs. *unsad), and for children to acquire positive counterparts of positive-negative linguistic items before negative ones, according to Boucher and Osgood (1969). Boucher and Osgood described the phenomenon in terms of an E(valuative)-factor, ascribing E+ values to positively-evaluated language use, and E- (negative) values to negatively-evaluated uses, such as the adjectival oppositions of good/bad, pretty-ugly, right/wrong, sweet/sour, and funny/sad. The same values will be applied to the analysis of proximatives in the empirical surveys below.

25Other studies illustrating the operation of evaluative choices in this way have included Akatsuka and Strauss (2000: 231) in their study of the use of counterfactual conditionals by survivors of the Los Angeles earthquake. Akatsuka and Strauss found that the majority of counterfactual conditionals used were in fact expressing counter-to-desire meanings at the same time. Kuteva et al (2014), in a comprehensive crosslinguistic study, also found a close association between proximity and counterfactuality and the aversion of undesirable situations. In addition, the Avertive ‘gram’ (grammatical function) of the proximative liketa (Kytö and Romaine 2005), is discussed as predicating undesirable situations in counterfactual complements in Southern US English dialects. In all such cases, it is hypothesised that the E-factor of Boucher and Osgood (1969) motivates the expression of aversion of unfavourable situations in proximative expressions, in a general preference for speakers to convert an assumed E- (negative) situation into an E+ (positive) one.

3. Diachronic data

3.1 Almost

26In Ziegeler (2010) the earlier history of almost was discussed; it will not be reviewed further in the present study except to point out that the adverb began life as a quantifier (+determiner), mæst ealle/ealle mæst, with the meaning roughly glossable as 'most all'. By Middle English the quantifier had spread to verbal predicates as a univerbation, and it is in verbal environments that counterfactual meanings, describing the non-occurrence of predicted events, began to emerge with past tense or past participle predicates, as indicated in the contrast between (7a) and (7b-c). However, only 7 out of 123 tokens of almost appeared in the Helsinki Corpus from 1450-1710 with verbal predicates (5.69%). Such frequencies, though low, are still fairly representative of the use of the adverb today. The number of tokens available from more recent corpora, on the other hand, is much higher than those available from the Helsinki Corpus, and so provides a more statistically-reliable estimation of the distribution of the proximative. The following table (from Ziegeler 2015, to appear) illustrates the spread of past tense and participle forms in the Corpus of Late Modern English Texts (Extended Version), CLMET(EV), the environments most likely to derive counterfactual interpretations, as discussed in 2.2.

  • 6 The selection included the first 30 files and the last 15 files of each sub-corpus, the corpora bei (...)

Table 1. Numbers of tokens of almost with bare preterite verb forms or with perfect or present participles appearing in a selection of files from CLMET(EV), shown as proportions of the total of verbal and non-verbal complements.6

Period searched


+ bare preterites


+ perf./pres participles




















27The number of preterite complements and present or present perfect participles was relatively small, indicating that it is possibly the case that counterfactual readings do not figure prominently in the usage of the adverb throughout the history of English, and that much of the debate on whether the negation meanings are entailments or pragmatic implicatures is largely exaggerated in its significance to the development of the functions of almost. However, the question remains how such counterfactual meanings arose in the first place, from an implicit orientational sense of proximative direction towards a specific target or goal. It was noticeable, at the same time, that of the cases of almost that were recorded in the corpus data, both historical and contemporary, a large number of examples were complemented by predicates referring to events which were either undesirable or unimaginable, i.e., evaluatively negative contexts (E-neg), leaving only a small number, in each survey, of predicates that could be evaluated as positive or neutral (E+), according to Boucher and Osgood's (1969) terms.

28The following data (also to appear in Ziegeler 2015) were obtained from the CLMET and CLMETEV corpora, revealing the overwhelming predominance of E-neg. contexts correlating with preterite and past participle verb forms.

Table 2. Frequency of evaluatively negative contexts in the use of counterfactual proximatives in the selected files from CLMET(EV).


almost + preterite/participles

E-neg contexts


CLMET 1710-1780




CLMET 1780-1850




CLMET 1850-1920




29A selection of the examples of preterite E-neg. contexts found in Ziegeler (2015, to appear) is provided below.

3.2 Examples

30E-neg or evaluatively-negative examples:


(10) Endur'd a sea that almost burst the deck. (1738, George Lillo, Marina)

(11) … he almost destroyed the whole colony: and then proceeded to Mona, with a resolution fully to complete the conquest of that island. (1760-1, Charlotte Lennox, The Lady's Museum)


(12) ... the real truth is, my poor mother has almost lost her senses … (CLMETEV: 1799, Frances Burney, Cecilia)

(13) This intelligence almost paralyzed me. (CLMETEV: 1839, Thomas Clarkson, The History of the Rise, Progress and Accomplishment of the Abolition of the African Slave-Trade, by the British Parliament)

(14) He almost fell when he went to open it. (1843, William Thackeray, Vanity Fair).


(15) Dearest Alick, The unlucky journey to Syria almost cost me my life. (1866-9, Lucie Duff Gordan, Letters from Egypt)

(16) Yet I doubted him—I almost called him cruel. (CLMETEV: 1850, Dinah M. Craik, Olive)

(17) There were moments when he almost lost faith in his whole system of education, ... (1918, Litton Strachey, Emminent Victorians).

31E+ or evaluatively positive/neutral examples included:


(18) … every body almost worshipped you … (Samuel Richardson, Clarissa Harlowe, or The History of a Young Lady).


(19) … ... Mountains, whose shaggy steeps appeared to be inaccessible, almost surrounded it. (1794, Ann Radcliffe, The Mysteries of Udolpho)


(20) The exquisite neatness of the room almost concealed its wretchedness. (1852, Chambers's Edinburgh Journal no. 418-462)

32Many of the examples with negatively evaluable complements also appeared to express clearly hyperbolic situations, as shown in (12), (13) and (14). Other examples of such uses included: almost deprived me of being, almost froze up the blood of Sophia, almost broken my heart, almost lost the use of all her faculties, almost cut into my heart, almost took away my breath, almost robb'd me of my understanding, almost drove me to despair, almost drowned me with Weeping, all appearing in the 1710-1780 corpus; almost broke our hearts, almost exceeded endurance, almost destroyed her, almost stopt his whizzen, almost disdained to breathe, almost paralyzed me, almost despised myself, almost subdued the feeble remains of her spirits, almost overcame her, almost gave him up for lost, all appearing in the 1780-1850 files, and almost unmanned me, almost intoxicated me, almost overwhelmed him, almost overpowered me, almost wept, almost killed me, almost began to despair, almost fainted, almost rent the drum of my ears, being amongst those found in the later, 1850-1925 corpus. Although expressing the avoidance of unpleasant circumstances, such exaggerations also lend a sense of unreality to the use of almost, thus contributing to the counterfactual meanings.

3.3 Nearly

33The history of nearly is not as long as that of almost, since the adverb did not start to appear until Early Modern English, being derived from the adjective or preposition near, and thus the earliest meanings simply meant 'closely', and the original adverb had no effect on the truth conditions of the complement with which it co-occurred. The Oxford English Dictionary places the first such appearance at 1540, found in contexts related to scrutiny and inspection. The first example of the adverb appearing in the entry for the OED Online: nearly is the following:

(21) If a man woll more nerelyer behold this thinge
'If a man will more closely behold this thing'
1540, R. Morison tr. J.L. Vives
Introd. Wyesdome (new ed.) sig. Bvj.

34Senses of physical proximity also feature in (21), and in (22):

(22) They pursuing the French men, in their fight did so nerely folow them into their holde ...
'They pursuing the French men, closely followed them into their hold in the fight ...'
1569, r. Grafton.
Chron. II. 98

35From 1561, meanings of physical proximity appear to develop into senses associated with blood ties or human or other relations:

(23) To be nearly acquainted with the people of different countries can happen to very few.
1753, Johnson,
Adventurer No. 84, 2.

36The last example with a similar meaning is cited as appearing in the late 20th century:

(24) He has in mind a variety of factors, some more nearly connected than others.
1984, S. James,
Content Social Explan. 95

37Relational meanings may be also associated with the basic sense meaning 'closely', and may only be differentiated by the contexts in which they appear. The adverb is also found to qualify verbs expressing similarity, e.g. resemble, the last citation dated at 1993:

(25) The ASU and the American Youth Congress (AYC) ... more nearly resembles mass organisations.
1993, N.Y. Rev. Bks. 13 Jan. 47/1.

38Nearly is also used in equative comparative clauses during the entire time span, as shown in the OED Online:

(26) The words and the purport, as nearly as I remember, were these.
1791, T. Paine,
Rights of Man, I. 124.

39And in the CLMET(EV) (see below):

(27) … but still she had an eye to it as nearly as circumstances would admit.
1748, S. Richardson, Clarissa.

40The more idiomatic comparative usage, not nearly + adj./equative cl., is dated from 1745 (e.g. not nearly so nice etc.). Interestingly, an older, hardbound edition of the OED (1939) dates the first use with the meaning of 'almost, not quite', as commencing at 1683, while the more recent OED Online brings the date forward to as late as 1936. However, the examples given in the online version refer only to an adjectival use, e.g. the nearly-dark, the nearly-kiss, with a meaning of 'not quite' but not occurring with a verbal complement, let alone a preterite verb most readily yielding to counterfactual senses. From the limited data supplied by the OED, it looks very much as though the only frequent verbal preterite or participle complements were those similar to (24), in which the meanings of the co-occurring verb or participle were synonymous with those of proximity in any case.

  • 7 Forms searched included all the available spellings found in the Oxford English Dictionary Online; (...)
  • 8 The nearly tables did not include present participles, as had the almost data, since the almost stu (...)

41A search of the Helsinki Corpus yielded only 2 examples of nearly in the period 1570-1640,7 and one example in the period 1640-1710, revealing its relative infrequency compared with almost at the time. However, the CLMET(EV) Corpus, being much larger, provided a vast number of tokens, which were far too great in number to all be analysed. For this reason, the search was reduced, as with almost in the tables above, to the first 30 files and the last 15 files of each time period. The results were classified according to the function and/or meaning of the adverb, and are provided in the following tables. In Table 3, as with almost, only the co-occurrences with past tense verbs or perfect and passive participles were included, these being considered the most likely environments in which counterfactual senses may start to appear, as shown above in the contrast between (7a) and (7b-c).8 Table 4 shows the correlations of E-negative (E-neg.) contexts with nearly followed by a preterite verb or a past participle, and Table 5 displays a breakdown of the environments, and the correlation of E-neg. contexts with either preterites or participles.

Table 3. Numbers of tokens of nearly with bare preterite verb forms or with perfect or passive participles appearing in a selection of files from CLMET(EV), shown as proportions of the total of verbal and non-verbal complements.

Period searched


+ preterites


+ participles




















Table 4. Frequency of evaluatively negative contexts with preterite or participle predicates in the use of proximative nearly in the selected files from CLMET(EV).


nearly + pret./participles

E-neg contexts


CLMET 1710-1780




CLMET 1780-1850




CLMET 1850-1920




Table 5. Frequency of evaluatively negative contexts in the use of counterfactual proximative nearly in the selected files from CLMET(EV), comparing the ratio of E-negs. in preterite to participle contexts.


E-neg contexts





CLMET 1710-1780






CLMET 1780-1850






CLMET 1850-1920






42The results will be discussed in the next section. Meanwhile, examples of E-neg. nearly appearing in the CLMET(EV) corpus included the following:


(28) ... the noble family, whose only disgrace is, that so very vile a man is so nearly related to them.
1748, Samuel Richardson,

(29) ... he pretty nearly exhausted both the Purse and the Patience of the valorious Plaintiff Walter Warmhouse.
1765-70, Henry Brooke, The Fool of Quality.

(30) ... and thou and CHESTERFIELD, with your united councils, had nearly led me to every-thing odious ..
1776, Samuel Jackson Pratt,
The Pupil of Pleasure.


(31) I was more than once nearly choaked with gall during the honey moon ...
1781, Richard Brinsley Sheridan,
The School for Scandal.

(32) She opened it, and, going forward, had nearly fallen down a steep, narrow stair-case ...
1794, Ann Radcliffe,
The Mysteries of Udolpho.

(33) ... he tried in vain to save them, which roused the indignation of the crowd, and nearly cost him his life.
1839, Charles Dickens,
Barnaby Rudge.


(34) ... When suddenly called by her own name, she nearly fainted with agitation.
1890, William Booth,
In Darkest England, and the Way Out.

(35) ... securing the anchor, and two of them were nearly drowned, though they swim like fish;
1866-9, Lucie Duff Gordon,
Letters from Egypt.

(36) That might be, but he very nearly threw me into a consumption ...
Chambers's Edinburgh Journal nº 418-462.

43E+ or evaluatively positive/neutral examples included:


(37) All my views are nearly completed, as far as regards external preparation.
1779, Samuel Jackson,
Shenstone Greeen.


(38) … until the effects of a stiff jungle-fever, that nearly made me proprietor of a landed property measuring six feet by two, sent me back …
Punch, Vol. 1


(39) I have found nearly everything that I want.
1910, E. M. Forster,
Howard’s End.

44Like the almost data, there were numerous examples which expressed proximity to E-neg. hyperbolic circumstances, though not all of these had verbal predicates, e.g., in the 1710-1780 corpus: nearly smothered him with their caresses, nearly exhausted, nearly in a State of Beggary, nearly suppressed by a gush of tears, nearly despairing of ever being able to find her, nearly blind to the way I went, nearly distracted with joy; in the 1780-1850 corpus: nearly choked with gall, nearly drowned her eyes with tears, nearly bewitched, nearly overcome with joy, nearly hugged themselves out of breath, nearly coughed himself into a fit of apoplexy, turned up her nose until she nearly fractured its bridge, nearly parboiled me in an over-heated bath, nearly killed by the intelligence, parting … nearly killed me, nearly sent Jos off his legs, … collars nearly choked him, nearly choked himself dead with laughter, nearly threw him into fits, nearly plunged him again into the abyss of despair, nearly squeezed the breath out of my body; and in the 1850-1920 corpus: nearly fainted with agitation, nearly died of laughing, nearly scared out of their wits, nearly in a state of suffocation, nearly distracted, nearly broke my heart, nearly mad with appetite, nearly demented, nearly stunned, nearly frightened into a fit of apoplexy, nearly threw me into a consumption, nearly senseless. However, the frequency of such uses was noticeably less than for the almost data, with nearly co-occurring in a greater variety of contexts, as shown in the tables, many of these being non-verbal and referring to measurements and quantities (e.g. nearly all). The presence of hyperbolic predicates was not in any way overwhelming, and many E-neg. predicates were not hyperbolic in nature (e.g. (32)).

4. Discussion of results

4.1 Almost

45In the earlier study of almost (Ziegeler 2015, to appear) it must be noted, the data demonstrating the co-occurrence of E-neg. contexts concerned mainly preterite verbs in co-occurrence with almost, since these formed the majority of uses (the proportion of participles was limited to perfect and present participles for the purposes of that study). This proportion concerned a maximum of only 15.2% (20/131) of such environments likely to be found in the 1710-1780 period, a maximum of only 10.7% for the middle period (1780-1850), and 15.4% for the later period (1850-1920). Thus the proportion of E-neg contexts for preterite environments would still have been extremely high, given the relatively small ratio of participles to preterites found in the data at the time.

46The conclusion of that study was to reveal that the correlation between preterite (bare past tense forms of the verb) and counterfactual readings of almost was dominated by an overwhelming number of examples referring to undesirable or unpleasant predictions that were averted (hence the use of the term Action Narrowly Averted by Kuteva 1998, 2001, to describe such grams crosslinguistically). A further conclusion was reached suggesting that the reason for the overwhelming tendency to associate counterfactual readings with the avoidance of undesirable circumstances was the need of the speaker to express intersubjective meanings. In some languages, proximity to a counterfactual situation is expressed by the use of a verb of desire, e.g. in Persian (Heine and Kuteva 2002, 313):

(40) mixast bemirad
      want.3sg.IMPF die.3sg.SUBJ.PRES

      He was about to die'.

which cannot mean (literally): 'he wanted to die', and yet it does convey the semantics of a prediction, rather than a negative entailment, since it cannot be known that he was about to die if he did not die at all. It may be questioned how a complete reversal of the meaning of positive prediction to negative prediction may occur: this takes place in the presence of 'switch-contexts' (Heine 2002), in which the original senses of volition or desire no longer make sense or are less probable. Possible reasons for the introduction of such meanings which are contrary to those most expected in such contexts are that the meanings of proximity compete for some time with the former lexical meanings of desire, and eventually win out, assuming the entire, focal meaning of the construction. At the same time, with almost and nearly there are no desire meanings involved, but nevertheless, as with volition verbs expressing proximity, once the proximative meanings take focus in the meaning of the adverb, the speaker finds the most need to exploit such meanings in referring to predicted events which are undesirable; the use of a proximative becomes all the more newsworthy if in fact it expresses the avoidance of an impending disaster. Otherwise, there is little reason to express what is known not to have occurred at all. A similar point was made in Kytö and Romaine (2005: 25), and the newsworthiness of the averted situation is even more accentuated in the frequent appearance of hyperbolic examples, such as those listed above: almost deprived me of being, almost froze up the blood of Sophia, almost broken my heart, almost lost the use of all her faculties, almost cut into my heart, almost took away my breath, almost robb'd me of my understanding. The role of intersubjectivity was claimed to be related to the possibility that the averted, undesirable events are those which the speaker would prefer not to mention, and that in expressing the avoidance of such events, he or she is accommodating the emotions of the addressee; this offers an extension of the notion of linguistic intersubjectivity (as defined by Traugott 2010, 2012) as formerly involving considerations mainly of the cognitive stance and the social status of the addressee. The manifestation of such accommodation is never more clearly apparent as in the 'pleonastic' negation in the proximative expressions of languages such as Spanish, Polish, Bulgarian, and Chinese, mentioned in 2.3, in which intersubjectivity is formally encoded.

  • 9 In the case of (semi)-modal verbs, as with proximatives, it is not always predictable which Quantit (...)

47At the same time, the mechanics of Gricean and Neo-Gricean Quantity implicatures are held to be responsible for the reversal of orientation towards P as orientation away from P. According to Ziegeler (2000, 2010, 2015 – to appear), speakers' reports of proximative past events are often more associated with those of aversion than proximity, since the outcomes of past situations are usually known, while at the same time, past proximity predicts what might have happened otherwise. Such competing meanings are governed by the interplay of Horn's (1984, 1989/2001) neo-Gricean, R-based and Q-based implicatures. R-based implicatures (involving Quantity 2 and Relevance maxims) permit prediction meanings to arise out of proximative senses, by allowing for inferences over and beyond what is already expressed in the utterance; for example, the use of the semi-modal verb be able to (P) in the past tense (She was able to solve the problem) carries the implicature that (P) was accomplished (She solved the problem). Q-based implicatures (involving Quantity 1 and Manner maxims), on the other hand, act to cancel the prediction implicatures reversing the polarity of the prediction to express aversion and counterfactuality instead (e.g., She wanted to solve the problem → ‘but she didn’t’).9 This is effected, as noted earlier, in the fact that Q-based (or Quantity 1) implicatures operate on the provision of maximal information in an utterance, so that anything not said is believed 'not the case' (see also Levinson's 2000 Q-heuristic); in other words, selecting X for the purpose of denying Y. As noted in 2.2, if a prediction is made about a past event, the maximality of information is understood as increased by the fact that the past is usually known, allowing for denial of anything not said. Proximative expressions about the past, then, reverse their polarity to express distance, not proximity, from the truth conditions of the predicate, since the addressee expects that the speaker's maximum information was only of the proximity of P, not P itself. While these mechanics may explain the pragmatic reasons for a switch in polarity, it is nevertheless obvious that the employment of proximative constructions in referring to the avoidance of undesirable circumstances is a significant motivation for the switch to occur as well; nobody wants proximity to unpleasant events, and in the interplay of Q-based and R-based implicatures, Q-based will take precedence under such conditions. It is at this point that changes in the truth conditions of the predicate become important.

4.2 Nearly

48In the case of nearly, the data also point to a tendency to exploit the linguistic environment of preterite or participle predicates in order to express the aversion of unfortunate circumstances, though the statistics are slightly reduced compared with almost. There is also a considerable difference in the type of linguistic environment: the majority of E-neg contexts with almost appear with bare preterite verbs, while with nearly, it seems that participle forms, at least initially, take priority (see Table 5), though the distribution balances out during the last period (1850-1920). The frequency of E-negative contexts also rises steadily during the periods, though not reaching the same figures as almost; this is most likely because nearly is a more recent adverb than almost, and, in many respects, now in competition with almost (though not entirely, as we have seen above). The E-neg. contexts rise from 29.7% in the earliest period, to 48.2%, and then to 68.5% in the period 1850-1920, a rise of around 20% for each time span, while for almost, there is a steady decline of around 6% at each time-period. The converse statistics demonstrate a possible, gradual sharing of functions as nearly generalises its distribution to verbal environments. It should also be remembered that the almost data were more concentrated on co-occurrence of E-neg contexts with bare preterites, since such verb forms were considered to be the most likely to yield counterfactuality. Also, it was noted in Ziegeler (2010, 2015 – to appear) that the first example of almost co-occurring with a preterite verb observed in the Helsinki Corpus was dated at the 15th century, indicating that counterfactual meanings could have emerged quite early in the history of almost.

49In the case of nearly, the counterfactual meanings (E-neg.) seem to emerge much later, in the late Modern English period, as shown by the present data. However, in the 1710-1780 period and the 1780-1850 period, there is a noticeably high number of participle forms, 32% (as against 6.09% preterite forms), in the 1710-1780 period, producing at the same time a relatively high number of E-neg. examples (82.1%, as against 17.8% with preterite verbs). That is, for the first time period, 82.1% of the E-neg. contexts are found with participles, and for the second time span, 1780-1850, 78.3% of the E-neg. contexts are found with participles. However, at the same time, 78.3% come from 84.3% of the total number of preterites and participles for that period (145/172). At the last period, the statistics balance out equally, and the numbers of preterites and participles co-occurring with nearly is roughly the same. What this means is that the presence of E-neg. contexts was dominated by the environment of perfect or passive participles, relative to that of preterite complements; in other words, such an environment was providing the principle context in which the development of E-neg. meanings must have emerged. The participle context seems to balance out with preterite contexts at the later stage of the data, indicating that the senses of counterfactuality had spread to other contexts, there being a general increase in preterites (27 to 49 in the 1850-1920 period) co-occurring with nearly in any case, and a general reduction in frequency of participles at the same time (from 145 to 59 in the latter period). In the case of almost, there is a higher frequency of preterite verb forms than participles with E-neg. meanings. It may be predicted that preterite forms will eventually take over the E-neg. contexts for nearly as well, since participles, being intermediate in many cases between adjectival and verbal items, represent more time-stable categories than bare past tense verbs. The pathway of generalisation for nearly, then, may be predicted to begin with spatial contexts, as in (22), spreading to contexts expressing abstract relations (e.g. 28), and then to contexts expressing temporal relations (e.g. 32).

50While this is not intended as a grammaticalisation study, since as noted earlier, adverbs are a questionable case for grammaticalisation, it is noted in Traugott and Trousdale (2014: Ch. 2a: 29) that the category of adverbs is a very mixed bag in English, being comprised of items close to open-class, lexical items, and others, particularly focus adverbs, being considered closer to grammatical items. It is also noted that Heine and Kuteva (2002) label the proximative crosslinguistically as an aspectual category, in that it comes to serve a grammatical function. It is possible to consider that the emergence of the counterfactual meanings could be affected by principles of grammaticalisation after all. Nearly is a very clear case of a manner adverb expressing originally lexical meanings similar to 'closely', developing in certain contexts a more abstract, aspectual meaning of proximity and indeed focus, later to yield counterfactual meanings. In particular, as in a grammaticalisation account, the role of frequency of a certain text type is important in setting the context for a semantic change to occur. It is unlikely that semantic change, or any kind of diachronic change for that matter, will take place in an environment of low frequency (see, e.g. Hopper and Traugott 2003:126-130), and grammaticalising items are usually found to have a higher than normal text frequency (Bybee 2004). In the same way it appears that the environment of participles co-occurring with E-negative cases of nearly was frequent enough to permit the kind of changes that would lead to counterfactual meanings being derived from the proximity meanings they expressed. In other languages, the development of proximatives has been accredited to a grammaticalisation (e.g. Heine 1994, Heine and Kuteva 2002), since they involve lexical sources such as full verbs meaning 'want' eventually expressing an aspectual category, according to Heine (2004), as in (40) above. It is only when the aspectual marker comes to be used with past time reference that counterfactual inferences start to emerge. The difference is that with such proximatives, there is a grammaticalisation to the proximative meaning from another source, a verb of desire, while with nearly, the proximative meaning is already part of the semantics of the adverb from the start. Nevertheless, the adverb is eventually affected in its semantic trajectory by the interpretation of (un)desirability in the complement as well, since complements predicting events and situations which are contrary to desire count amongst those most likely to receive a counterfactual interpretation of proximity.

  • 10 A total of 55 tokens of this semantic type (out of the 246 tokens of nearly) was counted in the fir (...)

51It may be asked why such changes could take place. It is necessary to consider first the fact that both proximative adverbs, almost, and nearly, co-occur most felicitously in gradable semantic predicates, e.g. (23-24), and that with ungradable predicates, a kind of grading is forced into the meaning which is not there logically in the first place, such as in (14) or (32). It is also noted in the data that many examples co-occur with participle predicates expressing synonymous meanings of proximity, e.g. related, allied, concerned, connected, etc. as well as non-participle verbs expressing similarity or proximity of appearance, e.g. resemble, imitate and some expressing spatial proximity such as approach, border, threaten etc. All such verbs already carry meanings of proximity, and so collocate quite readily with nearly, which is logical since the semantics of the verbs like those of the adverbs also express closeness.10 In the case of the verbs expressing human relations, many are represented in the form of participles, and so constituting the most frequent environment in which E-neg. contexts were found also. The 'switch context' described, for example, by Heine (2004) in which a change to grammaticalised meanings takes place, is a context of ambiguity in which both the former meaning of the developing grammatical item and the later meaning are both possible. It is hypothesised in the present study that the switch context to counterfactual meanings took place in a context in which the former meanings of 'closely' were present, but dispreferred in the semantic context. Given that the original meanings of nearly meant 'closely', a change took place in which a manner adverb simply referring to spatial relations was reinterpreted as referring to temporal proximity falling short of the expressed predicate, and so changing the meaning of degree of intensity or manner to one of degree of factuality in the predicate. In such cases, the truth conditions of the affected predicates were significantly altered, by means of Gricean Quantity 1 implicatures, to suggest that the proposition expressed by the predicate failed to occur.

52The reinterpretation may be seen in several examples from the earlier period. Example (28) (repeated below as (41)) expresses a human relation of closeness which is nevertheless undesirable:

(41) ... the noble family, whose only disgrace is, that so very vile a man is so nearly related to them.
1748, Samuel Richardson, Clarissa.

53(42) illustrates a similar context:

(42) This sentiment of pity is nearly allied to contempt, which is a species of dislike, with a mixture of pride.
1751, David Hume, An Enquiry Concerning the Principles of Morals.

54Another example in which a preterite verb appears, conveys a similar ambiguity:

(43) ... by public storms, which at that time very nearly threatened him.
1753, Theophilus Cibber,
The Lives of the Poets of Great Britain and Ireland.

  • 11 The change could equally have occurred in the context of synonymous preterite verbs as well, as (47 (...)

55In all such cases, nearly could be taken as ambiguous between meaning 'closely' and 'almost, not quite', and the preferred reading would be understood to express the avoidance of the unpleasant situation expressed in the predicate. While this shift may not necessarily have taken place in the context of participles alone,11 their frequency in the data renders the possibility of such a tendency highly likely: they can express both 'closely P', entailing P (as in (41), 'a man so closely related to them') and 'not (quite) P (as in (42): 'This sentiment of pity is almost allied to contempt'). Such ambiguities may be illustrated in the following diagrams, in which in Fig. 1, the relationship is interpretable either one of parallel proximity: nearly P = closely P = P, while in Fig. 2, nearly P = not quite P = 'not p'.

Fig. 1: 'closely P'

Fig. 1: 'closely P'

Fig. 2: 'almost P'

Fig. 2: 'almost P'

56The first use of nearly in Figure 1 could be said to describe alignment, in which 'closely P' entails P since the adverb only modifies the intensity of P, while P already exists; in Fig. 2 nearly P does not entail P since the adverb no longer simply modifies the intensity of P, but grades the degree to which P actually holds. In such cases, depending on the hearer's evaluation of how much of P is required to call it P, it could be argued in Figure 2 that P does not yet exist. In this way, then, spatial proximity (illustrated in Fig. 1) is reanalysed in different contexts as temporal proximity (illustrated in Fig. 2). The status of such assumptions is not at issue in the present study, since it is hypothesised that the negation of P in such a way is related to a pragmatic implicature (Horn's (1989/2001) Q-based or Grice's Quantity (1)). However, in the present study it could also be argued that the switch contexts for the development of counterfactual meaning in nearly were brought about through the reinterpretation of 'closely' meanings of parallel proximity, indicating alignment with a goal, as those of prospective proximity, indicating orientation towards a goal, as in Fig. 2, rather than expressing the entailment of P, where the predicate referred to something undesirable. In other words, there was a retraction of the senses of proximity in cases in which the goal approximated to was undesirable or dispreferred, giving way to meanings expressing distance, emerging via the focus functions of nearly.

57Also observed in the data appearing during the 1710-1780 period was the use of nearly not only to express prospective or close parallels with P, but orientation away from P, as in the following curious examples:

(44) ... of his burnt, and he and his wife nearly escaped; that he afterwards died in King-street
1753, Theophilus Cibber,
The Lives of the Poets of Great Britain and Ireland.

(45) In the year 1623, when Prince Charles nearly escaped being cast away in the road of ...
1753, Theophilus Cibber,
The Lives of the Poets of Great Britain and Ireland.

(46) ... and in going over the last, they very nearly escaped falling into the enemy's hands;
1753, Theophilus Cibber,
The Lives of the Poets of Great Britain and Ireland.

58Although all three examples come from the same text, they are, nevertheless, fully attested. In such cases, the meaning of 'closely' is clear, they cannot be understood as meaning 'almost', but rather than expressing orientation towards P (the event of escaping) the meanings are retrospective, indicating orientation away from P 'they closely/narrowly escaped'. Such uses can no longer be found, as evidenced in their incompatibility with a present-day reading, due to the fact that counterfactual nearly may well be said to dominate present-day uses; the reader may ask, did the subjects escape or didn't they? However, if the sense of nearly is 'closely', they are not unidiomatic. Thus, three different proximity orientations may be isolated for the period 1710-1920 for the use of nearly at this time: parallel and prospective (with a meaning 'closely', but with an emerging meaning of counterfactuality), as well as retrospective (now predicted to be obsolete, and probably infrequent at the time it was used), as illustrated below:

Fig. 3. Multi-orientational model of proximative functions of nearly in the early 18thc. texts surveyed from the CLMET(EV).

Fig. 3. Multi-orientational model of proximative functions of nearly in the early 18thc. texts surveyed from the CLMET(EV).

59Therefore, it is in the domain of parallel proximity that the change is likely to have taken place towards counterfactuality, since the latter domain is no longer plausible with verbal predicates in present-day usage. Such a development may be seen as one of gradual pragmatic change over time, and the following steps are hypothesised to have taken place:

  • i. Nearly expresses the manner of spatial or physical proximity to an action or event (e.g. 22);

  • ii. Nearly extends to express abstract proximity in human and other relations (e.g. 44);

  • iii.In certain adversative past tense contexts, the expression of proximity to an event is reinterpreted to stand for both the proximity and avoidance of it; i.e. nearly P > not-P, (e.g. 43)

  • iv.Nearly P now expresses counterfactuality as a focus adverb (focusing on X for the negation of Y).

60Unambiguously counterfactual meanings, associated with Stage iv, are shown in (47):

(47) ... the nurse, in the plenitude of her bounty, nearly parboiled me in an over-heated bath;
Punch Vol.1.

61In examples such as (47), the hyperbolic use of counterfactual nearly is clearly present, as with the examples of almost. Stage (iii) is similar to the 'switch-context' described by Heine (2004) in which the ambiguity present represents a stage in which a change takes place (as noted earlier) and this is often a context in which the former meanings of a grammaticalising item are incompatible with the context into which they have spread; in the case of proximatives, an undesirable event is incompatible with a state of proximity to it. Such shifts are frequently found in grammaticalisation accounts, as are the shifts from spatial to temporal domains. The third stage (iii) has also been described as a grammatical metonymy (the VIRTUALITY FOR NON-ACTUALITY metonymy described by Ziegeler (2000) (see also Panther and Thornburg 1999)), in which it referred to the development of counterfactuality in conditional clauses. In such metonymies, one construction (proximity) comes to stand for another (proximity + aversion), which is actually contiguous with it.


62The simultaneous investigation of both adverbs, almost and nearly, in the late Modern English texts of the CLMET(EV) corpora, have revealed a great deal about the development of counterfactual meanings in both. In the former case, the counterfactual meanings arose earlier than the Late Modern English times, as evidenced in Ziegeler (2010), though it was not until the last 300 years that the tendency began to appear in which they were found with predominantly E-negative predicates in past tense and past participle uses. The counterfactual uses increase mainly as a result of the speaker's desire to intersubjectively accommodate the emotions of the addressee in expressing the past aversion of undesirable circumstances. On the other hand, with nearly, the proximative meanings are already in place in Late Modern English, with nearly P entailing P in participle constructions in which it functions as a manner adverb expressing physical or abstract closeness to P, rather than the degree to which P holds, a later focus meaning arising through the grammaticalisation of nearly as a focus adverb. Nearly P can be seen to be associated with three different orientation types of proximity meanings during the Late Modern English times: parallel, in which proximity is aligned with the goal; prospective, in which proximity is oriented towards P; and retrospective, in which proximity is oriented away from P. It is only the first type that could have shifted its meaning to include prospective proximity as well, as a generalisation of meaning; the last type is no longer found, and the prospective (counterfactual) proximity type may be taken as the dominant meaning in today's English. Thus although it may be argued in the semantics literature that nearly P carries an entailment of not-P, this was not always the case, as in the diachronic data examples abound in which nearly P actually entails P. The change from a more content-based meaning to a more abstract meaning expressing proximative aspect and focus is arguable as a grammaticalisation, but it remains for further studies on proximatives and other adverbs to add more interesting comparisons of this type to the present data.

Top of page


Akatsuka, Noriko McCawley and Susan Strauss. “Counterfactual Reasoning and Desirability.” Cause - Condition - Concession - Contrast. Cognitive and Discourse Perspectives. Eds. Elizabeth Couper-Kuhlen, and Bernd Kortmann. Berlin: Mouton de Gruyter, 2000. 205-234.

Amaral, Patricia. The Meaning of Approximative Adverbs: Evidence from European Portuguese. Doctoral thesis, Ohio State University, 2007.

Atlas, Jay D. “Comparative Adjectives and Adverbials of Degree: An Introduction to Radically Radical Pragmatics”. Linguistics and Philosophy 7 (1984): 347–377.

Atlas, Jay D. Logic, Meaning and Conversation. New York: Oxford University Press, 2005.

Biq, Yung-O. “Metalinguistic Negation in Mandarin”. Journal of Chinese Linguistics 17 (1989): 75–94.

Boucher, Jerry, and Charles E. Osgood. “The Pollyanna Hypothesis.” Journal of Verbal Learning and Verbal Behavior 8 (1969): 1-8.

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

Bybee, Joan. “Mechanisms of Change in Grammaticization: The Role of Frequency”. Handbook of Historical Linguistics. Eds. Brian D. Joseph and Richard D. Janda. Oxford: Blackwell Reference Online, 2004., accessed 28 January 2010.

Grice, H. Paul. “Logic and Conversation”. Speech Acts [Syntax and Semantics 3]. Eds. Peter Cole and Jerry Morgan. New York: Academic Press, 1975. 113-128.

Heine, Bernd. “On the Genesis of Aspect in African Languages: The Proximative.” Proceedings of the Berkeley Linguistic Society 20 (1994): 35-46.

Heine, Bernd. On the Role of Context in Grammaticalization. New Reflections on Grammaticalization. Eds. Ilse Wischer and Gabriele Diewald. Amsterdam: John Benjamins, 2002. 83-101.

Heine, Bernd. Grammaticalization. Handbook of Historical Linguistics. Eds. Brian D. Joseph and Richard D. Janda. Oxford: Blackwell Reference Online, 2004. Accessed 28 January 2010.

Heine, Bernd and Tania Kuteva. World Lexicon of Grammaticalization. Cambridge: Cambridge University Press, 2002.

Hitzeman, Janet. “The Selectional Properties and Entailments of ‘Almost’. Proceedings of the Chicago Linguistic Society 28 (1992): 225–238.

Hopper, Paul J., and Elizabeth Closs Traugott. Grammaticalization. Cambridge: Cambridge University Press, 2003.

Horn, Laurence R. “Towards a New Taxonomy for Pragmatic Inference: Q-based and R-based Implicature”. Meaning, Form and Use in Context: Linguistic Applications. Ed. Deborah Schiffrin [GURT 84]. Washington, DC: Georgetown University Press, 1984. 11-42.

Horn, Laurence R. A Natural History of Negation. Chicago: Chicago University Press.
[2001 edition published by CSLI, Stanford], 1989[2001].

Horn, Laurence R. Assertoric Inertia and NPI Licensing. Proceedings of the Chicago Linguistic Society 38, Part II (The Panels), 2002: 55-82.

Horn, Laurence R. “Almost Forever.” Pragmatics and Autolexical Grammar. In Honor of Jerry Sadock. Ed. E. Yuasa et al. Amsterdam: John Benjamins, 2011. 3-21.

Kuteva, Tania. “On Identifying an Evasive Gram: Action Narrowly Averted.” Studies in Language 22 (1998): 113-160.

Kuteva, Tania. Auxiliation: An Enquiry into the Nature of Grammaticalization. Oxford: Oxford University Press, 2001.

Kuteva, Tania, Bas Aarts, Geri, Popova, and Anvita Abbi, forthc. “The Grammar of 'Counter-to-Fact'.” unpublished ms., 2014.

Kytö, Merja, and Suzanne Romaine. “We Had Like to Have Been Killed by Thunder and Lightning.” Journal of Historical Pragmatics 6 (2005): 1-35.

Levinson, Stephen C. Presumptive Meanings. The Theory of Generalized Conversational Implicature. Cambridge, MA: The MIT Press, 2000.

Li, Charles N. “A Functional Explanation for an Unexpected Case of Ambiguity (S or ~S).” Linguistic Studies Offered to Joseph Greenberg, vol. 3. Eds. A.M. Devine and Laurence D. Stephens. Saratoga, CA: Anma Libri & Co, 1976. 527-535.

Liberman “And yet.” Language Log, March 28, 2004. Accessed 18 September 2014.

Liberman, Mark. “Nearly no: a gnarly knot.” Language Log, June 16, 2007. Accessed 18 September 2014.

Nouwen, Rick. “Remarks on the Polar Orientation of Almost.” Linguistics in the Netherlands 2006. Eds. Jeroen van der Weijer and Bettelou Los. Amsterdam: John Benjamins, 2006. 162-173.

Panther, Klaus-Uwe and Linda Thornburg. “The Potentiality for Actuality Metonymy in English and Hungarian.” Metonymy in Language and Thought. Eds. Klaus-Uwe Panther and Günther Radden. Amsterdam: John Benjamins, 1999. 333-357.

Peyraube, Alain. « Les ‘approximatifs’ Chinoise: Chabuduo, jihu, chayidianr. Cahiers de Linguistique Asie Orientale 6 (1979) : 49–62.

Pons Bordería, Salvador and Schwenter, Scott. “Polar Meaning and ‘Expletive’ Negation in Approximative Adverbs. Spanish por poco (no).” Journal of Historical Pragmatics 6 (2005): 262-282.

Sadock, Jerrold M. “Almost.” Radical Pragmatics. Ed. Peter Cole. New York: Academic Press, 1981. 257-271.

Sadock, Jerold. “Nearly and Almost.” Language Log, 24 June 2007. Accessed 18 September 2014.

Sevi, Aldo. A Semantics for Almost and Barely. Masters Dissertation. Tel Aviv University, 1998.

Schwenter, Scott. “Discourse Context and Polysemy: Spanish Casi.” Romance Philology and Variation: Selected Papers from the 30th Linguistic Symposium on Romance Languages. Eds. Caroline Wiltshire and Joaquim Camps. Amsterdam and Philadelphia: John Benjamins, 2002. 161-175.

Traugott, Elizabeth Closs. “Revisiting Subjectification and Intersubjectification.” Subjectification, Intersubjectification, and Grammaticalization. Eds. Kristin Davidse, Lieven Vandelanotte, and Hubert Cuykens. Berlin: Mouton de Gruyter, 2010. 29-70.

Traugott, Elizabeth Closs. “Intersubjectification and Clause Periphery.” English Text Construction 5 (2012): 7-28.

Traugott, Elizabeth Closs, and Graeme Trousdale. Constructionalization and Constructional Changes. Oxford: Oxford Scholarship Online (, 2014.

Van Dongen, W.A. Almost and Nearly. Neophilologus 6 (1921): 184-207.

Wierzbicka, Anna. “Precision in Vagueness.” Journal of Pragmatics 10 (1986): 597-614.

Zhu, De Xi. “Shuo chaydian (’On chaydian’)”. Zhongguo Yuwen 87 (1959): 435.

Ziegeler, Debra P. “What Almost can Reveal about Counterfactual Inferences.” Journal of Pragmatics 32 (2000a): 1743-1776.

Ziegeler , Debra P. “Running the Gauntlet on the Proximatives Debate : a Response to Recent Challenges.” Journal of Pragmatics 42 (2010): 681-704.

Ziegeler, Debra P. “Intersubjectivity and the Diachronic Development of Counterfactual Almost.” Journal of Historical Pragmatics (to appear, 2015).

Sources used

CLMET(EV): The Corpus of Late Modern English Texts (Extended Version), compiled by Hendrik de Smet.

Helsinki Corpus: Helsinki Corpus of English Texts, Diachronic Part. Compiled by Matti Rissanen, Merja Kytö, Leena Kahlas-Tarkka, Matti Kilipiö, Saara Nevanlinna, Irma Taavitsainen, Terttu Nevalainen, Helena Raumolin-Brunberg. Available from The Oxford text Archive. ICAME International Computer Archive of Modern and Medieval English.

OED. The Shorter Oxford English Dictionary on Historical Principles. 2nd ed. revised and edited by C.T. Onions. Oxford: Clarendon. 1939.

OED: Oxford English Dictionary. 3nd edn., online version. ( 2002.

Top of page


1 The data and examples shown in sections 3.1 and 3.2 also appear in that study.

2 Consider also whose expected frequency over the span of text in question is almost zero, in which the direction of the proximity suggests a point just falling short of zero.

3 Horn (2002) adapts this scale from Schwenter (2002), but both studies, being primarily concerned with the negative meanings of almost and its equivalents, do not discuss at any length why the proximative component of meaning should be entailed and asserted. Proximity is clearly an uncancellable meaning component, so perhaps the status of the negative/polar component would not be so debatable if the proximative meanings were not both entailed and asserted. In (3-4) and elsewhere, the term inference is understood as a general term covering entailments and implicatures.

4 However, this type of scalar analysis may not necessary apply to all verbal predicates (for this reason, Ziegeler 2010 classes almost as a semi-conventionalised implicature).

5 In fact, Nouwen (2006) supplies adequate information to suggest contradictions between temporal orientation and the presumed polar orientation of almost, e.g., Travis was almost on time (p. 8), in which the logical polarity is negative but the temporal orientation is positive.

6 The selection included the first 30 files and the last 15 files of each sub-corpus, the corpora being too large to be able to undertake a complete survey.

7 Forms searched included all the available spellings found in the Oxford English Dictionary Online; i.e. nerely, neerely, neerly, nerly, nearely, nereli, nearelar, nerely, nearely and neerlye.

8 The nearly tables did not include present participles, as had the almost data, since the almost study showed that they were considered not as relevant to counterfactual uses as perfect or passive participles.

9 In the case of (semi)-modal verbs, as with proximatives, it is not always predictable which Quantity implicature, Q-based or R-based, comes into operation. Matters of context and prosody also influence the direction of orientation (see Horn 1989[2001]).

10 A total of 55 tokens of this semantic type (out of the 246 tokens of nearly) was counted in the first period searched (1710-1780); i.e. 22.3%.

11 The change could equally have occurred in the context of synonymous preterite verbs as well, as (47) demonstrates, but as noted, these were not as frequent as participles.

Top of page

List of illustrations

Title Fig. 1: 'closely P'
File image/png, 1.5k
Title Fig. 2: 'almost P'
File image/png, 1.4k
Title Fig. 3. Multi-orientational model of proximative functions of nearly in the early 18thc. texts surveyed from the CLMET(EV).
File image/png, 6.7k
Top of page


Electronic reference

Debra Ziegeler, « Calamities and Counterfactuals: A Historical View of Polarity Reversal », Anglophonia [Online], 19 | 2015, Online since 15 July 2015, connection on 27 June 2017. URL : ; DOI : 10.4000/anglophonia.520

Top of page

About the author

Debra Ziegeler

Top of page


Licence Creative Commons
Anglophonia – French Journal of English Linguistics est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Top of page
  • Logo Presses universitaires du Midi