I’m afraid I’ll have to stop now… Your time is up, I’m afraid. Corpus studies and the development of attitudinal markers [1]

Gabriella Mazzon
University of Innsbruck


Within historical pragmatics, particular attention has recently been devoted to items that signal attitudes or have discourse-orienting functions, and that thus highlight the variable rate of subjectivity or intersubjectivity over time in specific genres. The abundant studies on modal markers, interjections, and linkers are all examples of this. The present contribution exploits these new advances for the analysis of the development of the marker I’m afraid, trying to trace the paths through which it acquired its present functions and syntactic position, from matrix clause to parenthetical or comment clause. The development is similar to that of other attitudinal items such as I think or I guess, but the pragmatic uses are different. The few diachronic studies available use various sources, but do not appear to tackle the various phases of this development in a systematic fashion. The present study employs corpora from different times, and concentrates on speech-related genres, which naturally show a greater abundance of attitudinal markers. Results show that some functions of the expression go back to Late Middle English, while others are relatively recent. Thus, it is possible to suggest a timeline not only for syntactic developments, but also for the occurrence of I’m afraid in different text-types and in different interactional contexts. This timeline reveals how different pragmatic functions gradually emerge from texts in a time sequence, and the overlapping that is typical of the multi-functionality of many discourse markers.

1. Introduction

The case study presented here concerns the development of the pragmatic values of the expression I’m afraid. This development started with the emerging of the construction in late Middle English (henceforth ME) dialogues. When studying markers of stance in the N-Town plays (Mazzon 2009), I noticed the following examples, in which the expression seems to show signs of being in transition towards parenthetical status, due to syntactic peripheralisation in (1) and that-omission in (2):

(1) He wyl lese oure lawe I am right sore afrayd (23.19)
‘He will loosen our laws, I really fear’ [2]

(2) I am aferd there wyll be sum thing a-mys (441b.349)
‘I am afraid there will be something wrong’

These uses, which are not present in other cyclic drama collections, are only a handful, yet they seem to point to a rather developed stage in the spread of relatively “modern” uses of this expression. Modern use is evident in the syntactic placement of the expression as parenthetical, though the intensification present in example (1) is not allowed in Discourse Markers (DMs) proper. [3] Also, the meanings of the expression in both examples do not seem to refer literally to “fear”, but rather to the anticipation, announcement or warning about something unpleasant being reported, as is the case with modern use. [4]

Therefore, the aim of this paper is to trace the development of this expression in older texts, including the syntactic path leading to its parenthetical placing, and to find traces of its present pragmatic values. The developments in question will be tracked through various corpora, in order to ascertain 1) whether any continuity or overlapping can be traced through corpora representing different textual traditions; 2) whether this element can be connected to the speech-relation structure modelled by Culpeper and Kytö (2010) in their book on dialogic elements; and 3) to what extent the ways in which I’m afraid was used in past texts reflect functions that are considered typical of today’s usage in authentic conversational samples, not least in academic environment, as exemplified in the title of the paper.

The starting hypotheses were formed on the basis of previous studies on parentheticals and on comment clauses, in particular the fundamental contributions offered by Laurel Brinton (2007; 2008). Accordingly, the expectation was that 1) there would be a clear syntactic development, similar to that undergone by I think and I mean, involving syntactic fixation, fairly regular tense construction, and non-interruptibility; [5] 2) there would also be a transition from Matrix clause or dependent clause to parenthetical; and 3) there would be a gradual emergence of pragmatic values starting from the literal meaning, with an increase in the frequency of subjective and intersubjective values. These are developments that have been shown to be the case for other pragmatic markers; for example, many first-person clausal DMs arise in ME and anticipate parentheticals (Brinton 2010: 289).

I assumed that the presence of rare but clear examples in the N-Town plays would mean that the pragmatic values were already established in Late ME, and that there would be other examples of parenthetical use at that time, although this is hardly mentioned in the few studies available that deal with this expression – the development does not seem to have been studied in a systematic way so far. The incidental emergence of the above examples in a relatively homogeneous corpus, the N-Town plays, led to the exploration of other corpora, which were sifted for this expression in its various contexts. This meant thorough sampling of wide-ranging, non-text-type-specific corpora, such as the Middle English Compendium (MEC), and the Innsbruck Computer Archive of Machine-Readable English Texts (ICAMET), but also of more specifically speech-oriented corpora, such as the Corpus of English Dialogues (CED). Particular attention was also devoted to letters, since these tend to be rich in stance expression. Similar corpora from Modern English were also sampled, while Present-Day English (PDE) data was drawn from the British National Corpus (BNC) and from corpora available on the ICAME Corpus Collection on CD-ROM (2nd ed. 1999) (see Sources for a complete list of the corpora used).

2. Theory and Models

Parenthetical first-person DMs such as I think, I guess, I’m sorry and I’m afraid (called “anchored parentheticals” by Wichmann 2001: 179) are quoted by Brinton (2008: 2) among the most common comment clauses. Contrary to other types of DMs, they only start to arise in Early Modern English (EModE) and Late Modern English (LModE). They are often inserted for stylistic reasons and have pragmatic functions, and thus do not signal discourse disfluency, as is often claimed, for example, for I mean (Wichmann 2001: 186). These pragmatic functions are mainly subjective or intersubjective: they introduce comments or indicate stance, afterthought or repair (Brinton 2008: 5–9). In many cases, the formal pattern of development seems to be from matrix clause to parenthetical disjunct and from there to pragmatic marker. [6] The expressions have a wider scope, and are often placed in final position. This is the so-called “Matrix-Clause Hypothesis” (Brinton 2008: 35ff.). However, Brinton also emphasizes that the historical data point to alternative paths, as shown by the very case of I’m afraid, since the expression is rarely followed by that-clauses prior to its spread as a parenthetical (Akimoto 2002: 3; Brinton 2007; 2008: 39–40).

One such alternative is an explanation through the adjoined adverbial structure “as/so I think”, which appears also in Old English (OE) and ME in as I wene and other “cognitive” predicates. The structure appears to be relatively recent: it first appears in EModE in various first-person comment clauses that later could have just lost their connector, taking on the present structure. This explanation, put forth in Brinton (2008), seems the most plausible for other comment clauses, e.g. (I) pray you and please or (as) you know. A third path would be the construction to be afraid of + NP, which is very common at least from ME onwards. Structures with inversion, or with omission of the complement, would be the cases that provide the transition point towards the parenthetical structure (for the possible integration of the latter two types of structures in the grammar, see Blakemore 2006). Opinions differ as to what extent cases like I’m afraid involve decategorialisation, pragmaticalisation and/or lexicalization (favoured by Fischer 2007 in explaining the rise of parentheticals), and even as to the way these phenomena interact with each other. [7] Brinton (2008: 69) also emphasizes the importance of idiomatization, which will be mentioned again below.

Akimoto (2002) described the development of I’m afraid using the Oxford English Dictionary (OED) CD-ROM (2nd ed.) as his sample. He noticed an increase in that-deletion after the seventeenth century, especially in first-person contexts. From this, and other concurring evidence, he concluded that the expression has undergone decategorialisation and grammaticalisation, and that semantic bleaching started even earlier. This study is still frequently quoted because it concentrates on this expression, but it remains to be seen whether this account is confirmed by wider corpus study; this is the purpose of the present paper. Consider, for instance, that the OED-based analysis allows for a category “parenthetical” only in the seventeenth century, and that the examples brought by Akimoto only date from the nineteenth century onwards. The preliminary evidence that gave the impulse to the present study leads to a different conclusion. Tissari (2007: 63) suggests that I fear was more common in EModE in the functions today allocated to I’m afraid; however, this hypothesis was not verified here.

3. Development in Middle English data

The various forms connected to PDE afraid were searched for in the whole Middle English Compendium (MEC) collection. 511 tokens were found, of which only 16 were in the first person. The fifteenth-century section of the ICAMET letter corpus yielded four examples in the first person, and the ME part of the Helsinki Corpus of Historical English (HC) only one first-person example. [8] The verb affraien, from Old French, has two participial forms, the regular affraied and the more definitely adjectival aferde; however, both were always predicative, and were never found in attributive position (OED). The fact that the two forms were perceived as not totally overlapping is shown by the existence of doublets, as in examples (3) and (4). [9] The locution be afraid appears in all syntactic and morphologic variations in ME; in the first person, it appears with intensifiers and complements reinforcing its literal value of ‘being frightened’ (as in examples 5–7), which is not the first meaning given in the Middle English Dictionary (MED), where the earlier meanings of the verb are translated as ‘attack’ and ‘disturb’.

(3) ...“Ly wiþ me, for to day þow despousedest and weddest me. I am þy god Venus.” And he, aferde and affraied, ledde þat nyȝt wiþ oute slepe;…
‘“Lie with me, because today you espouse and wed me. I am your goddess Venus.” And he, afraid and frightened, spent that night without sleeping’
(Polychronicon, book 6, chapter 26, 14 c.; MEC)

(4) ... This wif was not afferd ne affraied. / Bot boldely sche seide and þat anon / Mary I diffie þat fals monk …
‘This woman was not afraid or frightened, but boldly she said, and quickly: – Marry, I defy that false monk…’
(Petworth ms of the Canterbury Tales, The Shipman’s Tale, 1589–1591, end 14c.; MEC)

(5) ...And thus wexe I withinne wroth, / That outward I am al affraied, / And so distempred and esmaied. / A thousand times on a day / …
‘And thus I grow angry inside, that am so afraid outside, and so ill-humoured and dismayed, a thousand times every day’
(Gower, Confessio Amantis, 3.56–3.59, late 14 c.; MEC)

(6) …And thus thenkende thoghtes fele, / I was out of mi swoune affraied,…
‘And thus thinking evil thoughts I was frightened to a swoon’
(Gower, Confessio Amantis, 8.2858–2860, late 14 c.; MEC)

(7) …But neuerthelese my wyt ys so thynne, / And also of Dethe I was so afrayed, / That hit ys oute where hyt went ynne…
‘But nevertheless my spirit was so weak, and also I was so afraid of death, that what went in is out’
(The Assembly of Gods, 286, ll. 1998–2000, early 14 c.; MEC)

The transition towards non-literal values is attested in some other ME texts, more or less contemporaneous to the N-Town plays. In these cases, there are still few signs of the parenthetical construction, although the omission of that is a step in that direction. In contrast, the meaning is already non-literal, as in examples (8)–(11), where it is clear that there is no reference to actual fear, but rather to a negative feeling about something that is going to happen, or that has already happened:

(8) …I am afferde off hyre þat she shall nott doo weell.
‘I am afraid about her that she shall not do well’.
(Paston Letters, John Paston II to John Paston III, late 15 c.; MEC)

(9) …Goo to Clarionas myn owen lady dere, / haue here this ryng, bere it here for me, / I am aferde I shall hir neuer see. / Tell ye hir soo in very certente /…
‘Go to Clarion my own dear lady, take this ring here, bring it to her for me. I am afraid I will never see her, tell her so in great certainty’
(Generydes, 6769–6772, mid-15 c.; MEC)

(10) …Allas madame said sire Bors / I am aferd he hath bytrayed hym self and vs alle
‘Alas, Madam, said Sir Bors, I am afraid he has betrayed himself and all of us’
(La Morte Darthur, 18, chap. 15, 15 c.; MEC)

(11) New Gyse. Ey, ey! yowr body ys full of Englysch Laten. I am aferde yt wyll brest.
‘Hey, hey! Your body is full of English-Latin, I am afraid it will burst!’
(Mankind. PP. 154.1 – 167.412, 15 c.; HC)

Also significant with regard to the developments considered here are a trickle of examples with dependent clauses introduced by lest (examples 12–13), while in another couple of cases the syntax is still of the “Matrix-clause” type, but there is an even more obvious transition towards the new values of indicating stance and assessment of a situation, as in example (14), where the “heavy task” is the voicing of an opinion.

(12) …helpe me / for here by in a slade are syxe theues that haue taken my lord and bounde hym / soo I am aferd lest they wyl slee hym…
‘Help me, because near here in a valley there are six thieves who have taken my lord and bound him, I am afraid lest they will slay him’
(La Morte Darthur, Book 7, chap. 5, 15 c.; MEC)

(13) …for cawse þei thynke that I am aferd lest I shuld haue it no longere; and as for that I pray ȝow tell them…
‘… because they think that I am afraid lest I should have it no longer, and as for that I pray you tell them…’
(Paston Letters, John Paston I to Harry Waryns: Draft, late 15 c.; MEC)

(14) …By my hede sayd kynge Arthur I am aferd syre Mellyagraunce has taken vpon hym a grete charge
‘By my head, said King Arthur, I am afraid Sir Meligrance has taken a heavy task upon himself’
(La Morte Darthur, Book 18, chapter 7, 15 c.; MEC)

(15) … / vnkundenesse, vnkunninge, vnclannesse, beon arerd / so þat harmes þei boden, as ich am aferd. / Some go up, and some go down:
‘bad feelings, ignorance, uncleanliness, are well established, so that they bring harm, as I fear. Some go up, and some go down’
(Short Religious Poems, Sins of our Time, end 14 c.; MEC)

Apart from examples (1–2), there is only one other example in which there is a clearer shift towards the parenthetical in accordance with the Matrix Clause path indicated by Brinton (2008) for other comment clauses. This example (15) shows an “afterthought” structure. But it must be kept in mind that it was found in a poem, and could represent a line-filler, like the N-Town example (1), although this does not necessarily diminish its significance for our investigation. However, example (15) is also significant if we go with the alternative hypothesis reported by Brinton (2007) about I mean, i.e. that parentheticals develop more from the cases in which the structure governs a phrasal clause than from those in which it is a matrix clause itself. In any case, the shift involves a loosening of syntactic bonds, and this is indeed what we witness, although the number of instances is too low to draw any conclusions. – Nonetheless, it is clear that these examples show no direct parallelism between form and function in development.

Four interesting examples are found in the Paston Letters in ICAMET (for the pragmatics of letters, see Palander-Collin 2010). Examples (16) and (17) show signs of meaning change, from the literal meaning to that of “having/communicating the feeling that something unpleasant has occurred/is about to occur”, where the reference is to events that have already occurred and cannot therefore be “feared”. They are thus similar to example (14) above.

(16) I am aferd that Jon of Sparham is so schyttyl wyttyd, that he wyl sett hys gode to morgage to Heydon, or to sum other of ywre gode frendys,…
‘I am afraid that Jon of Sparham is so easily frightened, that he will set his goods to mortgage to Heydon, or to some other of your good friends’
(PASTON2, Margaret Paston to John Paston, mid 15 c.; ICAMET)

(17) For yn trowthe y am aferde that Roberd Rad_clyff hathe deseyvyd me, for he laboryd to me dayly by my Lords comawndement off Warwyk,
‘Because in truth I am afraid that Robert Radcliff has deceived me, for he importuned me every day at my Lord’s command of Warwick’
(PASTON4, Sir John Felbrigge to John Paston, 2nd half 15 c.; ICAMET)

The main element that the majority of ME examples have in common seems to be, as identified by Tissari (2007: 61), the transition towards a meaning that is “increasingly based on the speaker’s subjective belief, state or attitude towards the proposition”. This transition, which also concerns I fear in EModE, mainly involves future (possible) events, but also shows attention to the hearer’s face, and therefore is used for politeness purposes. This is probably the path through which I’m afraid came to be associated with “regret in saying something unpleasant” even in connection with past events – the “fear” ceases to be associated with facts (especially with upcoming facts) and starts to be connected to the speaker-hearer relationship, thus increasing the (inter)subjectivity of the item. This ceases to have literal referential meaning (announcing that the speaker is afraid of something) and acquires interpersonal functions, as is to be expected in the process of pragmaticalisation.

4. Early Modern English

More abundant evidence emerges for the EModE period, and particularly for the last part. In order to make the data more manageable, the search was limited to co-occurrence of afraid with the first person singular, and all the other contexts in which afraid appears were sifted manually. The samples initially used for this part were mainly the sixteenth- and seventeenth-century materials in the Letter Corpus of ICAMET (four examples) and the Corpus of English Dialogues (CED, about 80 examples). Eight relevant examples were also found in the HC, while further evidence was provided by the Corpus of Early English Correspondence Sampler (CEECS) and by the Renaissance materials in the Perseus collection (see below). I will discuss syntactic and pragmatic developments in combination, but separately for each set of data.

The four examples found in the Letters corpus range from 1587 to 1698, and they are all cases in which I am afraid is preceded by a linker or other material and followed by a clause with that-omission, as in example (18). The meaning shift is perceptible. The latest example (19), from 1698, shows this most clearly, as it is a case of anticipating negative consequences for the recipient rather than for the “speaker”. This is also the example that shows a clearest transition towards parenthetical status.

(18) …and I pray to God the Scottish King do deceive me, but I am afraid he will not.
(Charles Howard to Walsyngham, 1587, Letter Corpus of ICAMET)

(19) And I cannot end without making it my request that you should for my sake take some care of yourself and let not your trouble overcome you, which I am afraid it will do if you don't strive against it.
(Lady Giffard to Lady Berkeley, 1698, Letter Corpus of ICAMET)

The CED showed abundant and varied results both in drama and in fiction, not only keeping to some extent the difference between afraid and affrighted as the participial verb form (there are still scattered instances of the active form of the verb), but also showing I am afraid in a variety of constructions. Here, again, the clearest cases of parenthetical uses are those attested in the later decades of EModE, – for instance example (20) from the eighteenth century, but also example (21) from 1647 – where the range of pragmatic contexts is expanded. A clearly more “modern” structure (although not parenthetical as such) is found in the cases in which the expression follows a linker and precedes a coordinate predicate without repetition of an overt subject, as in examples (22) and (23), when it precedes a non-clausal structure (example 24), or occurs between subject and predicate, as in example (25).

(20) “The Person”, resumed she, “whom I must teach how to acquire my good Opinion, will, I am afraid, hardly recompense me by his Docility in learning, for the Pains I should be at in instructing him.
(Charlotte Lennox, The Female Quixote; Or, The Adventures Of Arabella, 1752; CED) [10]

(21) Thra. Th’art a pertish thing:
And – I’m afraid, have beene distastfull to him:
I’m halfe afraid on’t Girle: – we must be wise;
By’s frown we fall Wench; by his favours rise.
(Anthony Brewer, The Covntrie Girle, 1647; CED)

(22) “…for I do not use to see such terrible Objects here; possibly those you do see”, said the Count, “are indeed more agreeable, but I am afraid have not that regard to your Honour, as I have.
(Aphra Behn, The Lucky Mistake, 1689; CED)

(23) “His Conditions are very bad, Madam”, returned the Gardener; “and I am afraid are such as will one Day prove the Ruin of Body and Soul too.”
(Charlotte. Lennox, The Female Quixote; Or, The Adventures Of Arabella, 1752; CED)

(24) I add moreover, if I have any sense of my Lords Disposition, I think if he had known any such thing, he would not have stood his being taken, or made his Application to the King in this manner, I am afraid not so suitable to his quality.
(The Arraignment, Tryal & Condemnation Of Algernon Sidney, 1683; CED)

(25) Scrub. Because he speaks English as if he had liv’d here all his Life; and tells Lies as if he had been a Traveller from his Cradle.
Arch. And this Priest, I’m afraid has converted the Affections of your Gipsey.
(George Farquhar, The Beaux Stratagem. A Comedy, 1707; CED)

Moreover, here we find some cases in which the expression is included in a stretch of reported speech, not necessarily in first-person form (examples 26 and 27) and in quoted speech (example 28). [11] In this respect, an especially interesting case is example (29), which makes the potential ambiguity between the literal meaning and the ‘anticipation-of-bad-news’ meaning explicit: the questioner elaborates on this potential, and demands to know whether the reported utterance includes a stance marker or hedge, or whether it can be considered factual to any extent. The meta-discursive value of these examples, especially of example (29), confirms the importance of also looking at non-first person occurrences of afraid.

(26) …all my Masters that had vsed to sell some of their papers so deare as they doe at the Lottery, walke from one end of the Office to the other sighing, and pitifully complaine and say their time of Angell-gathering, they are sore afraid, is at an end.
(The Star-Chamber Epitomized: Or A Dialogue Betweene Inquisition A Newes Smeller, And Christopher Cob-Web, 1641; CED)

(27) My Lord Borlace said, let them not be affraid, for upon my honour…and that he was sorrie for his cosin the Lord Macguire, (and that he # was afraid he was taken).
(The Vvhole Triall Of Connor Lord Macguire, 1645; CED)

(28) … possibly she might be sav’d; that upon this, in some Agony, she answer’d “I’m afraid I have destroy’d that which would have hanged that Villain; but here take this Key.”
(The Genuine Tryal At Large Of Mary Blandy, 1752; CED)

(29) Mr. G. My Lord, I think she said she heard it. And I have said several times to her, the Popish Plotters would be destroyed: but she answered, she was afraid the Nation would be destroyed first.
L. C. J. Did she say she was afraid of it, or, that the Nation would be destroyed first? I ask you once more, we must try People according to their Oaths. By the Oath you have taken, when you said you thought the Popish Plotters would be destroyed, what Answer did she make?
(The Triall Of Elizabeth Cellier, At The Kings-Bench-Barr, On Friday June The 11th. 1680; CED)

Among the cases in which the meaning seems the most detached from the literal one are example (30), in which the expression occurs in the context of a threat, and example (31), where it introduces an apology. Nevertheless, the most common cases are warnings, as in example (32), and ‘breaking bad news’ (example 33), which, as mentioned, is different from the literal meaning in that it refers to past events rather than to future ones that one fears;it is the oldest and most common of the non-literal values. New values found in the CED corpus are the ‘confronting’ ones, in the context of retorting or in indicating refusal or rejection, as in example (34).

(30) “Would you have me”, Madam, replied she, “go to his House? I am afraid the Marquis will hear of it.”
“My father”, replied Arabella, “can never be offended with me for doing a charitable action.”
(Charlotte Lennox, The Female Quixote, 1752; CED)

(31) Bev. Madam, your most Obedient – I am afraid I broke in upon your Rest last Night – ’twas very late before we parted; but ’twas your own Fault; I never saw you in such agreeable Humour.
(Richard Steele, The Conscious Lovers, 1723; CED)

(32) Ld. Froth I hope Mellefont will make a good Husband too.
Cynt. ‘Tis my Interest to believe he will, my Lord.
Ld. Froth D’e think he’ll Love you as well as I do my Wife? I’m afraid not.
Cynt. I believe he’ll Love me better.
(William Congreve, The Double-Dealer, 1694; CED)

(33) Pres. I am afraid you have received ill advise from others, possibly you know what is become of that Prisoner that spake of Counsel; we cannot assign you…
(The Severall Tryals Of Sir Henry Slingsby, 1658; CED)

(34) Pooh, pooh! that’s the old Story – You are so prejudic’d. –
TUKELY. I am afraid ’tis you who are prejudic’d, Madam; for if you will believe your own Eyes and Ears...
(David Garrick, The Male-Coquette, 1757; CED)

As some interesting Late ME examples had surfaced in letter samples, especially the Paston Letters (examples (16) and (17) above), the CEECS was also examined. The fifteen relevant examples found yielded some cases of a parenthetical structure interrupting the main SVO sequence, mostly from the seventeenth century (examples 35–37). The meaning range did not present many surprises, but we do find a case of “warning” in a letter from Queen Elizabeth (example 38), and some cases of “apology”, for instance, in the context of showing worry of having been tedious in one’s correspondence, as in examples (39) and (40). Some similar cases, as in examples (41) and (42), were also found in the HC, in drama from the end of the EModE period, suggesting a shift towards an even more metonymic use (which only retains the ‘I regret’ part of the meaning), nearly equivalent to ‘excuse me’ and ‘I’m sorry’ (Tissari 2007: 86).

(35) I hope, howsoeuer you be pleased to beare with ther audacitie towards your selfe, yet you wil not suffar a strange king receaue that indignitie at suche caterpilars hand, that, instede of fruit, I am affraid wil stuf your realme with venom.
(<Q RO1 1590 ELIZABETH1> CEECS) [12]

(36) I write thease from Hampton Court, wheare the King and Queene intend to stay till Alhollandtide, and whear I am afrayd I shall be for the most part attending; and thearfore, yf that part of my house which is at my own disposing…
(<Q COR 1628 T2MEAUTYS>; CEECS) [13]

(37) Commend my service to my Ladie Carlisle and Bedford, and all my frendes, and particularlie to my poore Cozen Carr and his wife, whose busines I am afraide may miscarie by reason of my absence from the courte, which greives me much.

(38) If any respect whatever make you neglect so expedient a worke, I am affraid your careles hide wil worke your unlooked danger.

(39) I am afrayd I have too much presumed upon yo=r= patience in soe tedious a narration, butt yo=r= goodness will beare with more faults then this…

(40) I am afraid I have beene over troublesome unto you in this matter, which makes mee be thus large with you….

(41) Madam, you have the prettiest Prospect in the World; you have Looking-glasses all round you – But I’m afraid I tire the Company.
Ber. Not at all. Pray go on.
(The Complete Works Of Sir John Vanbrugh, 1640–1710??; HC)

(42) – Ladies, I’ll take my Leave; I’m afraid I begin to grow troublesome with the length of my Visit.
Aman. Your Lordship’s too entertaining to grow troublesome any where.
(The Complete Works Of Sir John Vanbrugh, 1640–1710??; HC)

A sampling of Marlowe’s and Shakespeare’s plays in the Perseus collection of Renaissance materials, where the expression is not very common anyway (two and ten relevant examples respectively), yielded nothing unexpected, with the main meanings having to do with ‘softening’ strategies, concerning either the past or the future.

5. Late Modern English

In the eighteenth and nineteenth centuries there is a considerable expansion of DMs as they appear in texts, especially in reported dialogues within various text-types This testifies to their establishment as conversational tools with an increasingly subjective function It is possible that this is a consequence of the increasing need for a closer association between such conversational elements, i.e. certain language patterns, and social acceptability norms, with the rise of notions such as “propriety” in speech and “polite conversation”. Perhaps there was a greater need for printed texts to serve as “role models”, or representative examples of a wider spectrum of interactional stance markers.

The samples for this period (only first-person instances) include the early American texts of the Perseus collection (nineteenth-century American texts and the Richmond Times Dispatch from the nineteenth century), and the Corpus of Late Modern English Prose. These yielded 27, 24 and 14 examples, respectively. Other text collections consulted, for example, the writings of Lady Mary Montagu, and parts of the Modern English Collection at the University of Virginia Library (see Sources), yielded hardly any relevant or surprising instances. Among the latter, only some fictional works such as Fielding’s Tom Jones yielded many examples of parenthetical I’m afraid.

From a formal point of view, the nineteenth century is, as noted by Akimoto (2002), the time in which the expression becomes more and more fixed and non-interruptible, especially since the be contraction increases its prosodic and phonetic compactness. Apart from this, I’m afraid appears more and more as a parenthetical structure, taking various positions, as shown in examples (43)–(45). From a pragmatic point of view, the functions that emerge, apart from the old ones of ‘softening the breaking of unpleasant news’ and ‘attenuating the anticipation of undesired outcomes’, are the apologetic function, as in example (46), and the apparently new one of ‘mitigating a contradiction’, as in example (47). The latter example seems to show a further increase in intersubjectivity, since it relates more to the dialogue dynamics than to the facts referred to in the dialogue: it looks like a negative politeness strategy employed to reduce the conflictive potential of contradicting the interlocutor.

(43) It is not particularly good or particularly original but it may do. It will be pommade I am afraid, this novel – but it is that is it not which the many headed Beast demands?
(The Letters Of Ernest Dowson, 101. To Arthur Moore, 8 June 1890, Corpus of Late Modern English Prose)

(44) Anyhow I hope to have my story, now in hand, finished early this week – I am going to call it, I think, “Fin de Siecle”. Parts of it satisfy me more than most things I have done; parts I am afraid are exuberantly bad.
(The Letters Of Ernest Dowson, 101. To Arthur Moore, 8 June 1890; Corpus of Late Modern English Prose)

(45) … I must work day and night this week to make up arrears including the paper for the Economic Journal, of which I will send you the first part in a few days. It will be very bad, I am afraid. But if you will think it over, and improve it, it may do.
(The Letters Of Sidney And Beatrice Webb, Vol I., Sidney To Beatrice 23.5.1891; Corpus of Late Modern English Prose)

(46) …At last she too — in case it be a woman-notices the change in her friend's look, and she springs to her feet and says, with sincere but tardy contrition, “I am afraid I have tired you.” “Oh no,” says the patient; “not at all.” It is her last gasp for that morning; she can scarcely muster strength to say it; but let us be polite or die.
(Thomas Wentworth Higginson, Women and Men, chapter 45, 1888; Perseus)

(47) – A graduate of one of the Universities of Great Britain came to me, shaking and trembling. –He said he had “come to me as he would go to a physician.” I said, “You must stop drinking.” “I can’t.” “You will die.” “I am afraid I shall, if I give it up; I can’t.”
(Richmond Times Dispatch, Nov 16, 1860; Perseus)

6. Present-Day English

The corpora surveyed for contemporary usage were those included in the ICAME collection and the BNC online version. The numbers of tokens found in these corpora, broken down by functions, are shown in Table 1 below. Detailed figures are only provided for PDE samples because this was the only period in which substantial numbers of examples for many different types were found. What emerged immediately is that the parenthetical structure is not typical of the majority of occurrences of I’m afraid: in the majority of cases this was followed by a dependent clause, with or without that. [14] This structure showed cross-co-occurrence with the various pragmatic functions, suggesting that there is no correlation between the pragmaticalisation of the expression and a specific developmental path in syntax. Thus, there seems to be no gradual emerging of specific constructions in direct correlation to specific pragmatic values to support the hypothesis that the pragmaticalisation of I’m afraid happened in co-occurrence with development from structures other than a main clause.This leaves us with the Matrix-Clause Hypothesis as the most likely explanation for the development of this expression, contrary to other comment clauses functioning as DMs.

Literal 168 (8.04) 128 (47.9)
Softening 1055 (50.5) 61 (22.08)
Apologetic 407 (19.5) 47 (17.6)
Contradicting 182 (8.7) 16 (5.9)
Refusing 189 (8.8) 13 (4.8)
Rejecting 40 (1.9) -
Correcting 12 (0.7) -
Threatening 34 (1.8) 2 (0.7)
Total 2087 267

Table 1. Raw figures for different pragmatic functions of I’m afraid in the PDE corpora analysed. Absolute token numbers for specific functions are accompanied by percentages (in brackets).

As for the meanings of I’m afraid, there is continuity with the previous stages, in that the functions recorded so far were all present in contemporary corpora. However, there are functions that were more frequent or that emerge more clearly (see Table 1). The apologetic stance is conveyed quite often, also in fictional dialogue. The softening function remains predominant and is also made explicit, as in example (48), both in parenthetical and non-parenthetical uses; the high frequency of this function is also due to the fact that it applies to several context types. A new sub-case emerges, that is, the use as reply in I’m afraid so/I’m afraid not.This sub-case is to be structurally, but not functionally, distinguished from I’m afraid S + Aux (Neg), as in I’m afraid I did(n’t)- consider that this is different from other short replies such as I think so/I don’t think so. The differences lie in the fact that, on the one hand, the structure remains the same, that is, the fixation degree is higher (for example, there is no do-support), as attested by the fact that in spoken corpora there are cases of further reduction to afraid not and ‘fraid so (Tissari 2007: 71–72). On the other hand, the structure with think enables Neg-Raising in the negative, which allows for a higher level of softening, while the affirmative counterpart can be used to agree, but for I’m afraid, all such uses are cases in which the reply is perceived as unpleasant for the hearer, and therefore in need of softening. This function is most often shown in cases where the “unpleasant truth” is formulated by the interlocutor, who then asks for confirmation. In both the ICAME corpora and the BNC samples, I’m afraid so was several times more frequent than I’m afraid not. There are also cases in which the expression is used in the context of rejection, refusal, correction or contradiction (examples 48–53). These kinds of usage appeared most recently in our samples as compared to others, since it was first found in Modern English texts, as in example (47) above. The construction appears also in threats proper, as in example (54).

(48) … as to whether the textbooks they were using were suitable for what they had in mind . “I’m afraid,” he hedged repeatedly , “I can’t comment on publications.”
(The Morning Shadow Problem, 1986; Australian Corpus of English K19)

(49) Warned by the last example we decide to pronounce the compound as one word PUkenui. “Wrong again, I am afraid.” The pronunciation requires two word stresses…
(Biggs, Bruce. Maori Spelling, 1986; Wellington Corpus J32)

(50) Hanson said, “I want to see him.” “I’m afraid that’s impossible.” “Why don’t you pick up the phone and ask Flash?”
(Derek Lambert. The Banya, 1991; Freiburg-LOB N29)

(51) “I need a volunteer,” says Philip, and briefly explains the Shadow Scheme. “Not my cup of tea, I’m afraid,” says Rupert Sutcliffe...
(David Lodge, Nice work, 1988; BNC)

(52) “Perhaps you’d like to pose yourself, along with Tracey?” Vic Wilcox guffawed. “I’m afraid you’ve got it wrong, darling,” said Everthorpe,
(David Lodge, Nice work, 1988; BNC)

(53) …his eyes managed to look bluer and sharper than ever, like the ice-cold waters of a sunlit fiord. “Nevertheless I am afraid I must insist on seeing him, Miss Varna.”
(Janet Tanner, Folly's child, 1991; BNC)

(54) “Your SS soldiers, sir.” “If you don’t co-operate, I’m afraid you’ll be seeing a lot more of them.”
(Piers Falconer, War in high heels. 1993; BNC)

7. Conclusions

A wide-span corpus analysis did not have much to add to hypotheses on the syntactic development path of I’m afraid, since there is hardly any quantitative basis for disconfirming the Matrix-Clause Hypothesis: there is no indication of constructions other than main clauses having any significant frequency, since, as mentioned, types such as as I am afraid only occurred very rarely. The syntactic development of the expression is thus hypothesised as in Figure 1.

Matrix Clause > that-deletion > different & variable positioning > parenthetical = (Late ME, MEC onwards)

    ➢ Fixation, disallowing discontinuity (eg intensification) = (EModE, CED onwards)

    ➢ Attrition / phonological reduction (EModE)

    ➢ Use in short answers with so/not (EModE, CED onwards, but increasing later)

Figure 1. Suggested timeline for the syntactic development visible through corpus study of I’m afraid.

The question concerning the appearance of modern pragmatic values for this expression can now be answered in greater detail. Figure 2 outlines an approximate time-line, with the development of values going from ‘fearing’ to ‘regretting’ to the more intersubjective ‘apologising’ to the polite marking of ‘softening’ not only a negative piece of news but also a negative speech act such as giving a negative answer to a request, or proffering another potentially face-threatening, and therefore mostly dispreferred, second-pair part.

     “I fear that something bad will happen” (Middle English onwards, MEC) >
     “I fear to tell you that something unpleasant is the case/has happened/ is about to happen” (Middle English onwards, MEC) >
➢ Future > past (Late ME onwards)
➢ Bad news for me/us > bad news for you (Late ME onwards)
     “I regret having to tell you this” > “I’m sorry to tell you this” >
➢ … because it’s bad news
➢ … because it damages your face (EModE)
     “I apologise” (late 16c??, letters) >
     “I am sorry to contradict/reject/refuse” (ModE, CED onwards) >
     “I am (perfunctorily) softening a threat/a conflictive conversational move” (ModE, CED) >
➢ … thereby reinforcing its seriousness: “this is going to be bad…” (PDE, BNC)
     “I am (perfunctorily) softening a contradictory statement” (PDE)
➢ … thereby reinforcing its validity: “this is going to be bad because I say so” (PDE)

Figure 2. Suggested time-line for the semantic/pragmatic development as visible through corpus study. References are made to periods when tokens were found in all corpora examined for that period, otherwise specific corpora are mentioned.

As for tracing a timeline for the appearance of the expression in various text-types, the following trends were observed, although obviously only confined to the samples analysed: in constructed dialogue (i.e. fiction, drama), I’m afraid starts to appear from Late ME onwards, but is sporadic until Late ModE; the same seems to be the case with texts containing constructed speech (i.e. letters, non-fiction). In recorded speech, that is, according to Culpeper and Kytö’s speech-relation model, trials, depositions, and transcripts (for which we start to have more substantial records later), the expression appears only in EModE but is rare until PDE, as testified by the low numbers of examples found. A more detailed qualitative study on Modern and contemporary sources, taking dialogue dynamics into account, could further highlight the emerging of the ‘prohibition’ or ‘threatening’ contexts, in which the marker again refers mostly to future events, but seems to have an intensifying, rather than a softening effect, signalling ‘patronising’ stance or otherwise stressing the speaker’s power, as in examples (55)–(58).

(55) I’m afraid I must ask you to leave,” I said.
(BetteHowell, Dandelion days. 1991, BNC)

(56) Whatever reply I find for that one, for that one, obviously it is not going to be “I’m afraid you’re making an illegitimate leap from fact to value.”
(AC Graham, Reason and spontaneity. Non-fiction 1985, BNC)

(57) “But I’m in charge of G.W. Fashions. And I’m afraid that means that what I say goes.”
(StephanieHoward, Miracles can happen. 1992, BNC)

(58) Mr. Speaker Order. That is just it. I am afraid that we do not discuss in the Chamber applications for private notice.
(Hansard extracts 1991–1992, BNC)

The diachronic corpus analysis showed that there is a clear developmental path of increased subjectivity and intersubjectivity in this marker as it emerges in a variety of text-types. It must be remarked, however, that constructed dialogues (within both speech-purposed and writing-based texts that tend to increase their speech-like quality, that is, fiction, plays and correspondence) seem to be richer in occurrences of “pragmatically loaded” I’m afraid than recorded dialogues, since more relevant examples were found in literary works than in non-fiction. The latter cases mostly concerned transcriptions of dialogues (as in example 58) or autobiographical writing. These cases should therefore be considered “authorial” markers, elements of interactivity inserted of a text precisely to increase the proportion of speech-like elements with a stylistic intent (Culpeper and Kytö 2010: 371).

In no case before PDE, however, were the pragmatic values exemplified above a quantitative majority: I’m afraid of and I’m afraid that, expressing the literal meaning, are still very vital. In view of the synchronic coexistence of different constructions and different meanings, it would be appropriate to state that I’m afraid has undergone idiomaticisation besides other developments (such as decategorialisation through fixation; Brinton 2008: 244), which accounts for the shift from “I fear” to “you should fear”.


[1] This paper is a revised and enlarged version version of a talk given at the worshop on Developing Corpus Methodology for Historical Pragmatics within the Helsinki Corpus Festival. Thanks are due to the organisers, the very responsive audience, and to Stephan Giuliani and Svenja Grabner for help with corpus analysis and with the revision. I would like to thank the two anonymous reviewers for their useful suggestions, the usual disclaimer about remaining shortcomings applies.

[2] The modern versions of the ME examples are my own loose translations, unless otherwise specified.

[3] This category, and indeed the label of “discourse marker” itself, has been variously described and defined. I follow Brinton (2010: 285–287) in considering DMs a functional (not formal) class of items, syntactically and prosodically independent (and therefore optional), non-truth conditional and not characterised by propositional meaning, but endowed with a range of procedural and pragmatic meanings – usually with textual and/or interpersonal functions.

[4] This shift was also undergone by the expression I fear (Tissari 2007: 58).

[5] This would particularly apply to texts with mimesis of dialogue, since e.g. narrative texts are bound to contain a higher proportion of other tenses, for instance I was afraid, over a longer time span.

[6] This involves the deletion of linking that, and the frequent placement of the expression in absolute initial position, which leads to its being perceived as syntactically detached from the rest of the utterance.

[7] The terms refer respectively to the processes of losing morphosyntactic flexibility, acquiring pragmatic functions, and acquiring structural autonomy, often with reduction from phrase to single morpheme, i.e. with univerbation (Brinton 2010: 302–305).

[8] The search was first carried out on all occurrences in order to have an overview of all uses. Later the tokens co-occurring with first-person use were selected and analysed.

[9] The two forms appear to be synonyms, although affrayed retains a closer relation with the verb by conveying a causative shade of meaning.

[10] Some of the transcription codes in examples from CED have been simplified for ease of reading.

[11] On DMs in Free Indirect Speech as either reinforcing or distancing authorial strategy, see Blakemore (2009), where it is argued that these forms are introduced by authors to increase “a sense of mutuality between reader and character” (138), but also, when they represent the author’s voice, to create “a sense of ironic distance” (150).

[12] According to the notation of this corpus, filenames contain the date and the sender’s name.

[13] An anonymous reviewer correctly points out that this example is ambiguously parenthetical.

[14] The contracted form of the verb is used in the vast majority of cases, regardless of text-type; over 1,800 contracted forms were found vs. slightly over 300 uncontracted am.


ACE = Australian Corpus of English 1986. Compiled by Pam Peters. Available on the ICAME Corpus Collection CD-ROM.

BNC = British National Corpus. 2010. Online version available at http://corpus.byu.edu/bnc/ – BNCweb (CQP-Edition, version 4.3, 3/1/2010). Link straight to the David Lodge quote in example (52).

BROWN = A Standard Corpus of Present-Day Edited American English, for use with Digital Computers. 1964, 1971, 1979. Compiled by W. N. Francis and H. Kučera. Brown University. Providence, Rhode Island.

CED = Corpus of English Dialogues 1560–1760, 2006 Compiled under the supervision of Merja Kytö (Uppsala University) and Jonathan Culpeper (Lancaster University). Available in the Oxford Text Archive.

CEECS = Corpus of Early English Correspondence Sampler, 1998. Compiled by Terttu Nevalainen, Helena Raumolin-Brunberg, Jukka Keränen, Minna Nevala, Arja Nurmi and Minna Palander-Collin at the Department of English, University of Helsinki . Available on the ICAME Corpus Collection CD-ROM and in the Oxford Text Archive.

Corpus of Late Modern English Prose. 1994. Edited by David Denison. Available in the Oxford Text Archive.

FLOB = The Freiburg – LOB Corpus of British English. 1999. Compiled by Christian Mair. Available on the ICAME Corpus Collection CD-ROM.

HC = Helsinki Corpus of English Texts. 1991. Department of English, University of Helsinki. Compiled by Matti Rissanen (Project leader), Merja Kytö (Project secretary); Leena Kahlas-Tarkka, Matti Kilpiö (Old English); Saara Nevanlinna, Irma Taavitsainen (Middle English); Terttu Nevalainen, Helena Raumolin-Brunberg (Early Modern English)  Available on the ICAME Corpus Collection CD-ROM and in the Oxford Text Archive.

ICAME Corpora of PDE on CD-ROM, 2nd version, 1999. [See also individually listed corpora]

ICAMET = Innsbruck Computer Archive of Machine-Readable English Texts. 1999. Edited by Manfred Markus. Available on the ICAME Corpus Collection CD-ROM.

LLC = London-Lund Corpus of Spoken English. 1980. Compiled by Jan Svartvik. Available on the ICAME Corpus Collection CD-ROM.

LOB = The Lancaster-Oslo-Bergen Corpus, POS-tagged version. 1986. Compiled by Geoffrey Leech,  Stig Johansson, Roger Garside, and Knut Hofland. Available on the ICAME Corpus Collection CD-ROM.

MEC = Middle English Compendium. 2006. Available at http://quod.lib.umich.edu/m/mec/. Link directly to Short Religious Poems and Sins of our Time.

MED =  Middle English Dictionary.2001. Available at http://quod.lib.umich.edu/m/med/

Spector, Stephen (ed.) 1991. The N-Town Plays. (Early English Text Society S.S. 11–12.) Oxford: Oxford University Press.

Modern English Collection of the Electronic Text Collection at the University of Virginia Library. Available at http://search.lib.virginia.edu/catalog?f_inclusive[digital_collection_facet][Online+Manuscripts]=1&f_inclusive[digital_collection_facet][Online+Texts]=1&catalog_select=catalog&op=AND&sort_key=title

Perseus Digital Library. Edited by Gregory R. Crane. Tufts University. Available at http://www.perseus.tufts.edu. Link straight to Women and Men, chapter 45 and the occurance of “Great Britain” from example (47).

Prose and poetry of Lady Mary Wortley Montagu [Electronic resource] / transcribed by Richard Bear, available through the Oxford Text Archive, http://ota.ahds.ac.uk/headers/1891.xml

WC = Wellington Corpus of Spoken New Zealand English.1998. Compiled by Janet Holmes, Bernadette Vine and Gary Johnson. Available on the ICAME Corpus Collection CD-ROM.


Akimoto, Minoji. 2002. “On the grammaticalisation of the parenthetical I’m afraid”. Studies in English Historical Linguistics and Philology. A Festschrift for Akio Oizumi (Studies in English Medieval Language and Literature, 2.), ed. by Jacek Fisiak, 1–9. Wien: Peter Lang.

Blakemore, Diane. 2006. “Division of labour: the analysis of parentheticals”. Lingua 116: 1670–1687.

Blakemore, Diane. 2009. “Parentheticals and point of view in free indirect style.” Language and Literature 18: 129–153.

Brinton, Laurel. 2007. “The development of I mean: Implication for the study of historical pragmatics”. Methods in Historical Pragmatics. (Topics in English Linguistics, 52.), ed. by Susan M. Fitzmaurice & Irma Taavitsainen, 37–79. Berlin: Mouton.

Brinton, Laurel. 2008. The Comment Clause in English: Syntactic Origins and Pragmatic Development. Cambridge: Cambridge University Press.

Brinton, Laurel. 2010. “Discourse markers”. Historical Pragmatics. (Handbooks of Pragmatics, 8.), ed. by Andreas H. Jucker & Irma Taavitsainen, 285–317. Berlin: Mouton.

Culpeper, Jonathan and Kytö Merja. 2010. Early Modern English Dialogues: Spoken Interaction as Writing. Cambridge: Cambridge University Press.

Fischer, Olga. 2007. “The Development of English Parentheticals: a Case of Grammaticalisation?” Tracing English through Time. Explorations in Language Variation. (Austrian Studies, 95.), ed. by Ute Smit et al, 99–114. Wien: Braumüller.

Mazzon, Gabriella. 2009. Interactive Dialogue Sequences in Middle English Drama. (Pragmatics & Beyond New Series, 185.) Amsterdam & Philadelphia: John Benjamins.

Palander-Collin, Minna. 2010. “Correspondence”. Historical Pragmatics. (Handbooks of Pragmatics, 8.), ed. by Andreas H. Jucker & Irma Taavitsainen, 651–678. Berlin: Mouton.

Tissari, Heli. 2007. “Compressing emotion to politeness: on I fear and I’m afraid”. Change in Meaning and the Meaning of Change. (Memoires de la Societe Neophilologique de Helsinki LXXII.), ed. by Matti Rissanen, Marianna Hintikka, Leena Kahlas-Tarka & Rod McConchie, 57–90. Helsinki: Société Néophilologique.

Wichmann, Anna. 2001. “Spoken Parentheticals”. A Wealth of English. Studies in Honour of Göran Kjellmer. (Gothenburg Studies in English, 81.), ed. by Karin Aijmer, 177–193. Göteborg: Acta Universitatis Gothoburgensis.