Actually,theres more to it than meets the eye

By Larry Bradley,2014-11-13 13:16
15 views 0
Actually,theres more to it than meets the eye

    1Actually, there’s more to it than meets the eye


     The Hebrew University of Jerusalem

    This paper presents a reanalysis of the syntax, semantics, pragmatics and phonology of the word actually. It endeavours to show, among other things, that in addition to the distinction between Greenbaum‟s „disjunct‟ and his „adjunct‟, we seem to require here a distinction between two kinds of „disjunct‟: (i) a „propositional modifier‟ and (ii) a „discourse modifier‟ (or „discourse transition marker‟). An examination of constraints on the use of actually as a propositional modifier suggests the

    need for a feature „assertive/nonassertive‟, not just on root sentences but on all structures that denote propositions, whatever their syntactic form.

    1. Actuallyand actually1 2

This paper is about the word actually, considered from the point of view of syntax, semantics,

    pragmatics, and phonology. I hope to show that the word is interesting both in itself and in a broader perspective.

     Actually is much more frequent in speech than in writing. The proportion is ca. 10.3:1 in the corpora of British English compared by Aijmer (1986), and ca. 3.4:1 in the corpora of American English compared by Oh (2000). It is also considerably more frequent in British speech than in American speech (2.4:1 according to Aijmer‟s figures and 2.2:1 according to Oh‟s), though less frequent in British writing than in American writing (1:1.5). The OED1 entry for actually (which

    dates from the 1880‟s) does not have a single example of the word either in sentence-initial or in

    sentence-final position; and Partridge 1947 gives a quotation from Ronald Knox which suggests that the great popularity of actually in English speech was something new even in the second thquarter of the 20 century: „They found Victor Lethaby a tornado of well-bred apologies, all

    punctuated with an irritating repetition of the word “actually” - a habit of modern youth,

    particularly when he is lying‟. Taken together, these facts suggest the possibility that some uses of actually, more specifically of what will here be called actually, represent a fairly recent 1

    innovation originating in Britain.

     The first to distinguish two kinds of actually seems to have been Sidney Greenbaum. He

    distinguished between actually as what he called an „attitudinal disjunct‟ and actually as an 2„intensifying adjunct‟. I shall refer to them provisionally as actuallyand actually. They are 1 2

    instantiated in examples (1) and (2) below:

(1) A: How was the bus ride? B: Actually, we went by train. 1

    (2) This isn‟t (just) hearsay, you know; I [actually saw it with my own eyes]. 2

As regards the syntax, we can start by saying that actually can serve as adjunct to a sentence, 1

    but not as IC of a verb phrase, and that conversely, actuallycan serve as IC of a verb phrase, 2

    but not as adjunct to a sentence. („Adjunct to a sentence‟ is here used in a sense akin to that of Greenbaum‟s (and Quirk et al‟s) „disjunct‟. But it is defined more narrowly than their „disjunct‟ in one dimension, since medial occurrences are excluded (see 2.1 below), and more broadly in another, since it subsumes their „conjunct‟). As for the meaning, both items mark the utterance or

     1 I am grateful to Jonathan Ginzburg, Eddie Levenston, Anita Mittwoch, and Yael Ziv for their comments on earlier versions of this paper, to Eliza Kitis for the reference to Hickey 1991, to Larry Horn for the reference to Schwenter and Traugott, and to Bas Aarts and two anonymous reviewers for helpful criticisms. 2 Greenbaum 1969: 127-128, 141-144. See also Quirk et al. 1972: 439 ff, 511, 666, 674 f. Quirk et al.

    1985: 583 ff, 62 f. In Quirk et al. 1985, the term „attitudinal disjunct‟ has been replaced by „content disjunct‟, and intensifiers have been placed under the new heading of „subjuncts‟.

    97060308.doc English Language and Linguistics 5.1: 1-16


     3 and both are the part of the utterance with which they are associated as in some way corrective,non-truthconditional. However, actually, can be described as „strong‟, as opposed to actually, 21

    which is „mild‟. For actually, it is possible to distinguish two principal uses, (a) „scalar‟ and (b) 24„truth-insistent. When the speaker uses „scalar‟ actuallyin some phrase (as in (2) above), he is 2

    envisaging a scale of properties with the content of that phrase at the top, and implying that any property below it on the scale would make the expression too weak. The „truth-insistent‟ use of

    actually serves to contrast what is really so with what is only pretended or imagined, and seems 2

    to be most typically found occurring within the scope of relativization or negation. Here is a 5corpus example of „truth-insistent‟ actually: 2

(3) Speaker A: what‟s happened to your finger

     Speaker B: well |I‟ve t\/old a 'lot of people• that I |dropped an |Anglo-'Saxon cr\oss on it•

    but |what I \actually d/id• was sl\/am it• in a |car d\oor• (CEC 1.9, TUs 199-


    As for actually, this merely indicates that the speaker is asserting something that is incompatible 16with some other proposition which has been expressed or implied. It can be called „mild‟ (or

    conciliatory), because if the speaker uses it in correcting someone else, it serves to make the correction less calculated to offend, and if he is correcting something said or implied by himself, an element of apology seems to be involved. Actuallyhas been described by Ilson (in 1

    Greenbaum 1985: 179) as being used for „modestly polite contradiction or amplification‟. According to Hickey (1991: 369), the key notion is „surprise: actually[i.e. actually JT] both 1

    marks what is said as surprising and mitigates the surprise either by warning the addressee beforehand or apologising to him immediately afterwards‟. In addition, Hickey suggests (p. 371)

    that „some surprise markers (notably „Actually‟)‟ serve „also as signals (sincere or not) that the

    contradiction involved is only a partial or unimportant one‟. Similarly, according to Oh (2000: ??), „[w]hen used withface-threatening acts, actually functions as a pragmatic softener

    (i.e., as a face-saving strategy).

     Levinson (1983: 87f) places actually, together with but, therefore, in conclusion, on the

    contrary, still, however, anyway, etc. under the heading of markers of what he calls „discourse

    deixis‟, which „indicate, often in very complex ways, just how the utterance that contains them is a response to, or a continuation of, some portion of the prior discourse‟. Levinson suggests that such items can be accounted for in terms of the Gricean notion of „conventional implicatures‟, i.e.

    „non-truthconditional inferences not derived from superordinate pragmatic principles‟ (p. 127). However, a nonGricean approach, based on a different theory of implicatures, is adumbrated in Smith & Wilson (1979: 180), according to which actually is one of a group of items (also

    including anyway, after all, still, and well) which the speaker may use to indicate to the hearer

    how the speaker‟s utterance is intended to be relevant. This idea has been more fully developed in

    Blakemore (1987), which treats such items (now generally referred to as discourse markers)

     3 Except in the use of actuallyas a discourse transition marker (see 4. below). 1 4 For an additional use, see 2.2 below. 5 This and the other corpus examples are from Svartvik and Quirk, A corpus of English conversation,

    1980 (CEC), which represents part of the London-Lund Corpus, which is a simplified version of the spoken texts of the Survey of English Usage at University College London. TU stands for Tone Unit, „the basic prosodic unit‟ in the analysis on which the CEC transcription is based. The symbol „•‟ marks the end of a tone unit, and nuclear pitch movements are marked by „\‟(fall), „/‟ (rise), or „\/‟ (fall-rise)

    before the nuclear vowel. The prosodic notation of CEC has been simplified by the use of the symbol „|‟

    to represent both onsets and boosters on nonnuclear syllables. (The „onset‟ is on the first syllable in a tone unit that is made prominent by pitch or stress, and a „booster‟ marks a stressed syllable other than an onset that derives additional prominence from its pitch level). Onsets and boosters that coincide with nuclear syllables have been omitted. Stressed syllables without additional prominence are marked by „'‟. 6 No scale of properties is envisaged. In its nonscalarity, actuallydiffers not only from actuallybut 1 2a

    also from in fact. See Schwenter and Traugott (2000) on the scalar properties of in fact „as an additive

    discourse marker‟.

    97060308.doc English Language and Linguistics 5.1: 1-16


within Relevance Theory as expressions that have procedural (i.e. nonconceptual) meaning and 7 It is not part of the aim indicate constraints on the inferential phase of utterance understanding.

    of the present paper, however, to discuss the relative merits of different theories of pragmatics.

    2. Syntactic-semantic properties

    2.1 Actually: a marginal element in sentence structure 1

    From the point of view of syntactic-semantic structure, actuallymay be called a MARGINAL 1 8element. In addition to functioning as adjunct in the sentence, either initially or finally, it may be instantiated as a PARENTHETICAL, i.e. a syntactic node that does not function as head,

    complement, specifier, adjunct, filler, extraposition or anything else that the grammar may

    provide for in the phrase of which it is an immediate constituent.

     „Parenthetical‟ has been defined in this sense, within the framework of HPSG, in Taglicht

    (1998). „Filler‟ is used in this framework to designate the nonhead node in unbounded dependency structures: structures of topicalization, relative clauses and WH-interrogatives (see Pollard and Sag 1994). „Extraposition‟ has been set up as an additional type of node in HPSG in Keller 9(1995) and Taglicht (1998).

     From a semantic point of view, a parenthetical relates to the sentence in which it is included like an adjunct to its head. Compare the different but equivalent versions of (4)-(6) with actuallyas 1 10parenthetical and as (pre- or post-) adjunct in different sorts of headed structure:

    . (4) you‟ve |gone 'over the 'M thr/ee• -

     a.. you |actually go \under it• (CEC 1.11, TU 1059)

     b. \/actually• you go \under it•

     c. you go \under it /actually

In (4a), actuallyis a Parenthetical node interposed between the Subject node (you) and its Head 1

    node (the VP go under it). In (4b) and (4c), it is an Adjunct node whose Head is the S (NP+VP) node you go under it, the order being AdjunctHead in (4b) and HeadAdjunct in (4c). The

    present paper will not deal with the significance of the choice between the two placements of actually as adjunct to the sentence. (This matter has been studied by Clift (1999), using the 1

    methods of Conversation Analysis.)

(5) a. This one, actually, I rather like.

     b. This one I rather like, actually.

In (5a), we see the Parenthetical actuallyinterposed between Filler (This one) and Head (I rather 1

    like) in a topicalization structure, and in (5b) it is the Adjunct following the sentential Head.

     (6) a. |I think the \awful thing /actually• a|bout 'getting eng/aged• is |having all your 'married

    fr\/iends a'round• who |tell you about 'married l\ife• . (CEC 2.10, TUs 1106-1109)

     b. |I think \/actually• the \awful thing a'bout 'getting eng/aged• is |having all your 'married

    fr\/iends a'round• who |tell you about 'married l\ife•

     7See also Wilson & Sperber 1993 and the references there. 8 „Marginal‟ is equivalent to the union of Quirk et al.‟s „disjunct‟ with their „conjunct‟. 9 In a theory that allows discontinuous constituents (McCawley 1982, 1989), marginal actually would be

    a sentence adjunct everywhere. On discontinuity in GPSG, see Borsley (1996: 230-234). 10.In the CEC prosodic transcription , „-„ denotes a „unit pause‟ (of one stress unit or „foot‟) and „‟ a

    .„brief pause‟ (of one light syllable), so that the sequence „- „ indicates something slightly longer than

    one unit pause.

    97060308.doc English Language and Linguistics 5.1: 1-16


    In (6a), there is an example of the Parenthetical between Head and Extraposition in NP structure,

     as Adjunct before its Head (the embedded clause the awful and in (6b) we have actually111thing…married life). A comparison of the sentences in (4)-(6) shows that the parenthetical placement of actuallydoes not affect its interpretation in relation to the sentence, though of 1

    course there are differences between the alternatives in terms of information structure and in terms of the relation of the sentence to the wider context.

     The marginality of actually is also reflected in its necessary exclusion from the scope of 1

    negation. Compare the three variants of (7):

     (7) a. Actually, no one had any objection.

     b. No one had any objection, actually.

     c. No one, actually, had any objection.

The relationship between actually and negation is the same in a, b, and c. 1

     It is worth noting here that the parenthetical status of actually(as of parentheticals in general, 1

    in the sense here intended) does not depend on its being separated by intonation boundaries from what precedes or follows it. This is exemplified by (4a) and (6a) above. In (4a) there is no boundary either before or after the parenthetical, while in (6a) the parenthetical is prosodically attached to what precedes it and separated from what follows.

    2.2 Actually: an integral element in sentence structure 2

Actuallyis an INTEGRAL element in sentence structure. It is never a sentence adjunct or a 2

    parenthetical, but is always the initial element of a sentence constituent. For example:

(8) This isn‟t just hearsay; I [actually saw it with my own eyes]. (VP) (=(2) above) 2

    (9) All these substances are harmful, and some of them are [actually poisonous]. (AP) 2

    (10) All these substances are harmful, and one is [actually a poison] (NP) 2

    (11) Of course what he‟s doing is scandalous, but is it [actually against the law]? (PP) 2

     (12) The ointment was even found to contain some [actually poisonous] substances. (AP) 2

It seems that any major phrasal category can be host to actually, provided that it is semantically 2

    appropriate. A phrase is appropriate as a host for actually only if it represents a SEMANTIC 2

    PREDICATE. APs qualify regardless of whether their syntactic function is syntactically „predicative‟, as in (9), or „nonpredicative‟, in the sense of being „attributive‟, as in (12). NPs and

    PPs, on the other hand, are predicates in their syntactically „predicative‟ functions, but not otherwise. Contrast (10) and (11) above with (13)-(14):

     (13) *He hadn‟t committed [actually a crime]. 2

     (14) *He wasn‟t accused [actually of a crime]. 2

Any attempt to place actually at the end of its phrase produces a sentence that English speakers 2

    will either treat as ill-formed or interpret as containing actually. For example: 1

     (15) a. You can be sure of one thing: if he [actually promises], he‟ll do it. 2

     b. *You can be sure of one thing: if he [promises actually], he‟ll do it.

     (16) a. Naturally, the people [actually at risk] are taking a different view. 2

     b. *Naturally, the people [at risk actually] are taking a different view.

     11 The same words in the same order could also be interpreted with actually in the matrix clause, but 1

    this would require actuallyto be prosodically subordinated to its Head (I think). 97060308.doc English Language and Linguistics 5.1: 1-16


     is not an option for the b. sentences, and they are accordingly ill-Here the reading with actually1

    formed. We may compare

     (17) a. He should have cut out the parts [not [actually essential to his argument]]. 2

     b. He should have cut out the parts not essential to his argument actually.

     (18) A: I don‟t expect he said that. You can‟t believe everything you see in the papers.

     B: a. I‟m not going by the press reports. I [actually heard the speech myself]. 2

     b. I‟m not going by the press reports. I heard the speech myself actually.

Sentence (17a), with scalar actually, might be a reply to the question „With so little time, what 2

    do you think he should he have done?‟, while (17b) might be a response to „Of course it was hard to follow, but with so little time, he had to speak quickly, didn‟t he.‟ Actually in (17b) in such a

    context would be understood as actually, and it would imply „I‟m afraid I disagree with you‟. In 1

    (18), too, we get a clear change in meaning. In B:a, actually is the scalar actually („I heard the 2

    speech myself, and you can‟t do better than that!‟), whereas in B:b it could be intended to forestall some other interpretation of the speaker‟s own preceding sentence, such as the inference that B did not have to rely on the press reports because he had access to a transcript of the recording. Here too actuallywould taken as actually. On the readings suggested, it would of 1

    course be natural for an editor to insert commas after argument in (17b) and after myself in


     While the marginal item actually is always outside the scope of negation, regardless of linear 1

    order, the integral item actually is always inside the scope of a preceding negator. For example: 2

(19) \no one• had |actually 'disob\eyed any 'order. (scalar actually) 22

    (20) well I |don‟t \actually 'do it like th/is• (CEC 1.4, TU 713) („truth-insistent‟ actually ) 22

     Though integral actually (actually), like its marginal counterpart, is defined in syntactic-2

    semantic terms, the distinction between 2a (scalar) and 2b („truth-insistent‟) correlates with a

    prosodic difference. 2a is always less prominent prosodically than its head, while 2b is typically 12more prominent (as exemplified in (19) and (20)).

     There is a third use of integral actually (actually), which occurs only in WH-2

    interrogatives. We shall classify it as 2c and call it „WH-insistent‟. It is prosodically similar to

    2a (the scalar use), in that here, as in 2a, actually is less prominent than its head, but its 2

    meaning is different. Though syntactically integral, like 2a and 2b, it relates semantically to the content of the question, insisting on the need for the information sought and rejecting the inference that might be drawn from the preceding discourse that the relevant information has already been given. For example:

     (21) A: What kind of job has he got? - B: He works for IBM. - A: Yes, but what does he

    actually do? 2

     (22) The reason that your work was not cited in the paper…probably has to do with the

    fact you never (correct me if I'm wrong) ever explain in detail how this wonderful 13parser you have actually works.(Internet discussion, 8 Oct 1998) 2

     12 2a, unlike 2b, can be totally lacking in intonational prominence, as in (i):

    (i) 573 and the |first f\/our•

     574 are the |only 'people who are 'actually 'going to m\ake it• - (CEC 2.9)

     13 If we substitute a yes/no interrogative, the actually will be interpreted as type 2a (scalar): 2

97060308.doc English Language and Linguistics 5.1: 1-16


    The syntactically integral status of 2c is clear from its inability to stand in initial or final position like marginal actually. Consider the inappropriateness in the above contexts of

     (23) …A: Yes, but actually, what does he do?

     (24) …how this wonderful parser works, actually.

    3. Syntactic structure and intonation

     and actually has consequences for the intonational The syntactic difference between actually12

    phrasing options. Compare the following two sentences (with „•‟, as before, marking the end of an

    Intonational Phrase):

     (25) the \house /actually• be|longed to her \parents• (not to her) 1

    (26) the \/house |actually be\longed to her 'parents (they weren‟t just renting it) 2

The attachment of actually to the house on the level of intonational phrasing in (25) above is in 1

    accordance with a general option for parentheticals to be prosodically attached in this way to the preceding, though not to the following node. Consider the following examples, with parenthetical 14however and you see:

     (27) a. the \house how/ever• be|longed to her \parents• (not to her)

     b. *the house• however belonged to her parents• (not to her)

     c. the \house you /see• be|longed to her \parents• (not to her)

     d.*the house• you see belonged to her parents• (not to her)

    We may contrast this with the intonational phrasing of adverbs that are syntactically dependent on the following VP. For example, with rather and completely as VP-initial adverbs:

(28) a. \/some of them• rather \liked the 'speech•

     b. but |nearly \/all of them• com|pletely |misunder\stood what he 'said•

However, utterances containing medial actually are not necessarily disambiguated by the

    prosody. Thus (4a), repeated here as (16),

     (29) you |actually go \under it•

can only be read as containing actually in the context in which it occurs (see above); but an 1

    identical utterance occurring in a different context could be interpreted as containing scalar actually. 2

(i) …it is not clear whether this wonderful parser you have actually works. 2

     14 The rule that parentheticals can be grouped with the preceding but not with the following node follows from a more general rule that applies to all medial nodes in nonbinary syntactic phrases. See Taglicht 1998.

    97060308.doc English Language and Linguistics 5.1: 1-16


    ): propositional modifier and discourse modifier 4. Two uses of marginal actually (actually1

    4.1 Actually in questions 1

We have assumed so far that marginal actually always points to the rejection of some

    proposition previously expressed or implied in favour of the one asserted by the sentence to which it belongs. So we might not expect this item to occur in questions, as it is not the function of questions to make assertions. Yet we have sentences like

(30) a. Was he actually Canadian, then (and not American)? 1

     b. Was he Canadian, then, actually (and not American)? 1

These can be paraphrased as in (31):

(31) a. Am I right in inferring that actually he was Canadian (and not American)? 1

     b. Am I right in inferring that he was Canadian, actually (and not American)? 1

    The primary illocutionary force of the sentences in (30) is interrogative, but they contain an implicit tentative assertion (or suggestion); and this, it seems, is sufficient to make it a suitable host for actually in the sense we have considered so far. The implicit tentative assertion is 1

    signalled in (30) by the presence of then, which indicates that the speaker‟s question is prompted by something he has inferred. The same effect can be produced by the use of initial so instead of

    final then:

     (32) a. So was he actually Canadian (and not American)? 1

     b. So was he Canadian, actually (and not American)? 1

    But the presence of an implicit tentative assertion can also be indicated in other ways. Compare

     (33) Did you (actually) want something else? and 1

     (34) Did you (*actually) want anything else? 1

The positive bias of (33) (due to the presence of some) allows it to accommodate this use of

    actually, whereas the neutrality of the question in (34) (signalled by any) makes the actually sound odd.(34) is overtly marked by any as lacking the sense of tentative assertion (or

    suggestion, or supposition) without which a question is not a suitable host for this use of 15actually. 1

     But there is also the possibility of using actuallyin a different sense. We can have, for 1


(35) Actually, was he Canadian? 1

    Here we are not concerned with any assertion, implicit or explicit. Here the speaker indicates that he has only just realized that the question is relevant, and that he is now about to go ahead and ask it. We may say that here actually serves as a DISCOURSE MODIFIER („discourse transition 1

    marker‟ would be more precise but clumsier), whereas what we had in previous instances may be

     15 There is no such constraint on actually. See 2

     (i) Of course what he‟s doing is scandalous, but is it [actually against the law]? (=[11] in 2.2 2


    (ii) …but is any of it [actually against the law]? 2

97060308.doc English Language and Linguistics 5.1: 1-16


    PROPOSITIONAL MODIFIER. If we compare (35) above with (36) below, we have a called a

    minimal pair:

     (36) Was he actually Canadian? 1

with actually as a discourse modifier in (35), as against actually as a propositional modifier in 1 1


    4.2 The discourse modifier in statements

    The distinction between propositional modifier and discourse modifier can also be made in statements. When actually is used to show that the speaker is „going off at a tangent‟, that a thought has just come into his head which changes the direction of the discourse, at least momentarily, we have the discourse modifier. So the mention of a person, say John, may prompt 16an utterance like (37):

(37) ACTUALLY, we must have a TALK about John. (discourse modifier) 1

    The same sentence, differently accented, may occur without any previous mention of John, or of the need for any kind of talk (more of less „out of the blue‟):

(38) ACTUALLY, we must have a TALK about JOHN. 1

Such changes in direction may also be signalled by the use of items like incidentally, by the way,

    or (when prompted by the preceding discourse) that reminds me. The last of these is not

    infrequently combined with actually, as in (39):

(39) ACTUALLY, that REMINDS me, we must have a TALK about John. 1

The following extract from a surreptitiously recorded CEC text (2.5, TUs 557-580) has two 17instances of actually marking a change in the direction of the discourse:


     opts [m] opts opt as a verb has been - - around - in my life as long as I remember

     |h/\as it•


     +it‟s |been+ a'round in m\y life•

    .@ but [m] I |think I‟m a'ware of it as a com'parative n\ovelty•

    .@ [m] |ie I |don‟t think it was a'round in my p\/arents‟ 'life•

     well |opt \/out•

     |must be a a |very pe'culiar c\/oinage 'actually

     |meaning *|not to take \/any 'of [dhi:]•

     *?the optative 5 to 6 sylls?*

     |normal \options•*

    . |y\es• -

     16 In [37] - [39], capitalization of a word indicates that it bears an intonational nucleus. 17 In accordance with the practice of the UCL Survey of English Usage, the words of Speaker have been left without prosodic transcription, because he (unlike and ) knew that the conversation was being recorded.

     There are a few symbols here that have not previously been encountered in the examples: „@‟ stands for schwa, „"‟ for heavy stress, and „*yes*‟ or „+yes+‟ for simultaneous talk. „{yes}‟ marks a subordinate TU.

    97060308.doc English Language and Linguistics 5.1: 1-16


     +to do |nothing ex'cept z\ero•+

     +?the optative syll?+

     |I sus{p\ect to 'opt is a} b\ack-for|mation• - -


     |actually *a a*n\other thing I‟ve |learnt most |interestingly from this |Open University

    I‟ve been |following this |spr/ing•

     |[/m]• - -

     |th\/is 'letter•

     which |stands for 'us for - - z/ero• -

    . |y\es•

    . |now \I‟ve 'always |taken th/at as a•

     |Gr\eek letter•

     but it |\isn‟t•

This part of he conversation begins by addressing the question how long the verb opt has been in

    use. When Speaker C decides not to go on with this or with the previous topic - the derivation of the verb opt - and to revert to something mentioned just before that, namely the phrasal verb opt

    out, he uses sentence-final actually (well |opt \/out• |must be a a |very pe'culiar c\/oinage

    'actually). When Speaker B returns to his view (already expressed earlier) on the derivation of opt, C ignores B‟s utterance and broaches a totally unconnected topic, the linguistic symbol for zero. This time he uses initial actually (|actually *a a*n\other thing I’ve |learnt most

    |interestingly from this |Open University I’ve been |following this |spr/ing•). An additional

    instance of actually as a discourse modifier occurs in CEC 2.10, TUs 1106-1109, given as example (6a) above, and repeated here as (41):

     (41) |I think the \awful thing /actually• a|bout 'getting eng/aged• is |having all your 'married

    fr\/iends a'round• who |tell you about 'married l\ife•

    Here the speaker has just been entertaining the company by talking about an embarrassing but funny mistake she made some time ago in a conversation, in the presence of her future husband, at a time when they were engaged. She now drops this incident and takes up the topic of engagements.

     Perhaps we should recognise the propositional modifier and the discourse modifier as distinct

    lexical items. At any rate, the one points to an incompatibility between the proposition asserted and some other proposition, stated or taken to be understood, while the other is a signal given by the speaker that he is initiating a distinct move, which deflects the discourse from the path it was 18following. While the term „corrective‟ may be appropriately used to characterize the meaning of actually, as a propositional modifier, this will hardly do for the discourse modifier, and it may be 1 19better to follow Clift (1999) in calling the latter a marker of „topic movement‟.

     18 As pointed out by Bas Aarts, actually as discourse modifier, unlike actually as propositional 11

    modifier, cannot be replaced by in actual fact. The latter is quite rare in conversation (only two instances in CEC), and is typically contentious in tone, rather than mild or apologetic, but it resembles actually 1

    as discourse modifier in pointing to an incompatibility between propositions.. 19 See also Oh (2000 ??): „Actually can…introduc[e] a new topic, orshiftthe focus or perspective on

    the current topic (Tognini-Bonelli, 1993:205; Lenk, 1998:174-182)‟. Clift (1999) distinguishes between

    „topic shift‟, associated with initial actually, and „topic change‟, associated with final actually. The

    former is said to mark „movement within a topic‟, and latter „movement from one topic to another‟ (p. 43). However, this „is intended to be neither predictive nor prescriptive (p. 44), and in fact the instances in example [40] above do not accord with the generalization.

     In Clift‟s analysis, position is described in relation to TCUs (Turn Construction Units), not

    sentences, but „sentence-initial‟ and „sentence-final‟ seem to entail „TCU-initial‟ and „TCU-final‟


     Halliday and Hasan (1976: 239 ff.) distinguish between „internal‟ and „external‟ conjunctive relations, where „internal‟ means „inherent in the communication process, in the forms of interaction

    97060308.doc English Language and Linguistics 5.1: 1-16


    4.3 Assertiveness

Sentence (33) above, repeated here as

     want something else? (42) Did you actually1

    is an example of what Bolinger (1977: 31) calls a conducive question. Such questions he considers to be „subtypes of suppositions.‟ Quirk et al. (1985: 808) similarly distinguish between

    conducive questions and questions with „neutral polarity‟. Conducive questions may have positive or negative orientation (or bias), and positive bias is marked by what they call assertive forms (some, some-, always, already, still, etc), which are distinguished both from corresponding non-

    assertive forms (any etc.) and from negative ones (no etc.). I suggest that we need to set up a

    pragmatic feature „assertive/nonassertive‟, such that only in the presence of the feature value

    „assertive‟ will it be natural to use actually as a propositional modifier, but that this requirement 1

    does not hold good for actually as a discourse modifier, or of course for actually. Assertiveness 12

    is not just the property inherent in assertions, in the sense of the term assertion as used in speech act theory. Sentence (42) has the feature value „assertive‟ without being an „assertion‟, and the same is true of

     (43) Was he actually Canadian, then (and not American)? (=(30a) above) 1

    Neither (42) nor (43) are „indirect speech acts‟ as this term is usually understood, though there is a similarity. In „indirect speech acts‟, we typically have the use of a grammatical structure primarily associated with one kind of illocutionary force, e.g. a yes/no question, for the purpose of conveying a different kind of force, e.g. a request. For example in a sentence like

     (44) Would you mind closing the door?

we might say that „question‟ is the direct or overt speech act, or literal illocutionary force, and

    that there is, in addition, an indirect or covert speech act, an illocutionary force that has to be inferred by the hearer, namely „request‟. The indirect speech act represents the speaker‟s real intention and the direct serves as a disguise. An appropriate reaction, therefore, must include a response to the indirect illocutionary force. It‟s not sufficient to answer

     (45) I wouldn‟t mind at all

and leave the door open. With a sentence like (46) (=(43))

     (46) Was he actually Canadian, then (and not American)? 1

a response to the literal illocutionary force alone is adequate:

     (47) Yes, he was.

But it is also possible to say

     (48) You‟re right, he was

    between speaker and hearer‟, and „external‟ means „inherent in the phenomena that language is used to talk about‟. But H&H make no distinction that would separate the „discourse modifier‟ from the „propositional modifier‟. Both are „internal‟ in their scheme.

    97060308.doc English Language and Linguistics 5.1: 1-16

Report this document

For any questions or suggestions please email