Chapter 3 Substitution 3-I Substitution and ellipsis In tllls and the next chapter we mall be discussing another type of cohesive relation. which takes two different forms: substitution~ and ellipsis. These can be thought of in simplest terms as processes within the text: substitu- tion as the replacem.ent of one item by another, and ellipsis as the omission of an item. Essentially the two are the same process; ellipsis can be inter- preted as that form of substitution in which the item is replaced. by nothing. But the mechanisms involved in the two are rather different, and also, at least in the case of ellipsis, fairly complex; so we shall devote a chapter to each. 3·'·' SllhstUvtiDn and refmnce The distinction between substitution and reference is that substitution is a relation in the wording rather than in the meaning. It has been empha- sized already that the classification of cohesive relations into different types should not be seen as implying a rigid division into v..-aterright compartments. There are many instances of cohesive forms which lie on the borderline between two types and could be interpreted as one or the other. The situation is a familiar one in many fields, and when one is attempting to explain phenomena as complex as those of human language it would be surprising to find things otherwise; this is particu- larly so when we are concemed with phenomena which are both semantic and gr.nnm.atical, since it frequently happens that semantic criteria suggest one interpretation while grammatical criteria suggest another, and the description has to account for bo~ facing both ways at once. The analysis that is adopted here is based on certain general principles. to which particular instances can be more or less tmambiguously referred.
J.I SUBSTITUTION AND ELLIPSIS 89 The principle distinguishing reference from substitution is reasonably clear. Substitution is a relation between linguistic items, such as words or phrases; whereas reference is a relation between meanings. In terms of the linguistic system. reference is a relation on the semantic level, whereas substitution is a relation on the lexicogram.m.atical level, the level of granunar and VQcabulary, or linguistic 'form". Ellipsis, as we have already remarked, is in this respect simply a kind of substitution; it can be defined as substitution by zero. So we have: Type of cohesive relation: Linguistic level: Reference Semantic Sub~titurion (including Ellipsis} Grammatical The meaning of the reference item he is 'some person (male), other than the speaker or addressee, who can he identified by recourse tQ the environment'. The cohesion lies in the semantic identity; and the fact that in a given instance the relevant environment may be the preceding text, in which, say, John Smith has occurred. is incidental Anaphoric reference, as we have seen, is merely a special case of reference in general. and the text is merely a special case of the environment; the reference may just as well be exophoric, where the relevant environment is the situation. Anaphoric and exophoric reference are both derived from the general underlying notion of recoverability of meanings from the environment. Substitution, on the other hand, is a relation within the text. A sub- stitute is a sort of counter which is used in place of the repetition of a particular item. Fo.r example. in [3; I] a. My axe is too blunt. I must get a sharper one. b. You think Joan already knows? - I think everybody does. one and tkes are both substitutes: one substitutes fOr axe, and does for kncws. And whe:reas in reference there is no implication that the pre- supposed item coul.:! itself have figured in the text:. and in many instances we know it could not have done, this IS implied in the case of substitution. Thus.. in [3 :I a and b] it would be entirely possible to • replace' one by axe arul dO£s by~. It £011mvs that. as a general rule, the substitute item has the same structural fimction as that for which it substitutes. In the above example,
90 one and axe are both Head in the nominal group; and does and knows are both Head in the Yerbal group. The i.dentity is less obvious in (3:2]: [3 :2] Has. Barbara !eft?- I think so. where the substitute so stands. for (that) Barbara has left. But here too the so has the same function in rdation to I think as has a dausc of reported spee-ch. Again, we have seen that reference is different; there i.s no such restriction there, and the grammatical function of a reference item may be quite different from that of its referent (example [2: 21) and [2:22]). From the point of view of textual cohesion, of course. substitution resembles reference in being potentially anaphoric, and hence constituting a link between parts of a text. But here too there is. a difference, following from the different nature of the two types of relation. Because reference is basically a non-verbal relation, a reference item may point in any direc- tion, and pointing to the preceding text is only one among the set of possibilities. Substitution, on the other: hand, being a verbal relation. is essentially confined to the text. Exophoric substitution is fairly rare; and it has the effect of implying that something HAS been said before. If the ftsherman sees me admiring his catch, he may say, without my having uttered a word: [3 :3] Ah! but you should have seen the one that got away. In doing so., however, he 'puts into my mouth' some such observation as That's a good-sized trout you've got there. [ myself might even have said That's a good-sized one you've got there, using exophoric substitution, in the first place; even here, howeYer, there would be a shared assumption that the fish in front of us was already the topic of conversation. The vast majority of aU instanCeS of substitution ate cndophoric; and of these again, the vast majority are anaphoric, although we shaH come across the possibility of cataphoric substitution under certain circumstances. Nearly every occurrence of a substitute, in other words, is a source of cohesion with what has gone before. 3.1.2 Types of substitution Since substitution is a grammatical relation. a relation in the wording rather than in the meaning. the different type5 of substitution are defined. gr.amm.atically rather than semantically. The criterion is the grammaricaJ function of the substitute item. In English. the substitute may function as a noun. as a verb, or as a clause. To these correspond the three types of substitution; nominal. verbal, and clausal.
3.2- NOMINAL SUBSTITUTION 91 These will be dim!Ssed in turn: nominal substitution in 3.2, verbal substitution in 3-·3 and clausal substitution in 3 .4. The following is a list of the items that occur as substitutes; the list is very short: Nominal: one, ones; sa~ Verbal: Jq Clausal: so, net There are .1. few expressions in which there is some indeterminacy among the three types. for example, do so, do the same; these will come up for discussion where they seem most appropriate. In addition, there is a borderline where substitution shades into lexical cohesion, involving the use of GENERAL WORDS such as thing in a cohesive flDlction. For the discussion of these see the chapter on lexical cohesion (Chapter 6}. 3.2 Nominal substitution The substitute one/ones always functions as Head of a nominal group, and can substitute only for an item which is itself Head of a nominal group. Fore~ple: (3 :4] I shootthe hippopotamus With bullets made of platinum Because ifi use leaden ones His hide is sure to flatten 'em.* Here bullets is Head of the nominal group bullm made of platinum and ones is Head of the nominal group leaden ones. The t\Vo nominal groups need not themselves have the same function in the clause; either may have any function that is open to a nominal group. Sometimes, as v.':ith reference, the presupposed item is huried deep inside a complex structure: the hearer gencrally has no difficulty in recovering it {g'[2:23} above); [3 : 5] lf only I could remember where it was that: I saw someone putting away the box with those candles in I could finish the decorations now. -You mean the little coloured ones? The substitute may differ from the presupposed item in number; in the following the presupposed item is the Mngular cherry. whereas the substi- tute is plural: [3 '6] Cherry ripe, cherry ripe, ripe I cry. Full and fair ones.- come and huy. * H. Bclloc, 'The Hippopotamus • in ~ Bml Chil.t's B<wk of :&asts, Duck-worth.
92 But the noun that is presupposed is always a count noun; there is no substitute form for mass nouns. Contrast [3:73 and b]: [3: 7) a. These biscuits are stale. - Get some fresh ones. b. This bread's stale. -Get some fresh.. In (b) the only possible form of substitution is substitution hy zero~ which 1s what we call ellipsis (Chapter 4). Semantically, ellipsis and substitution are very dose; we have said that ellipsis can be interpreted as substitu- tion without a substitute. Grammatically~ however, the two are fairly di<tinct. Some further examples of tme/ ones as substitute : [3: 8] a. So she wandered o~ talking to herself as she went, till on turning a sharp corner, she came upon two &t litde men,. so suddenly that she could nat help starting back:, hut in another moment she recovered herself, feeling sure that they must be- TWEEDI.BDUM AND I WEBD!.EDHH ••• They stood so still that she quite forgot they were alive. and she was just looking around to see if the word '1 WWIDLB' was written on the back of each collar, when she was startled hy a voice coming fro-m the one, marked • DUM'. b. I've heard some strange stories in my time. But this one was perhaps the strangest one of all. c. Which kind of engines do you want? Ones with whistles, or ones without? d. My dear, I trally must get a tlrinner pencil. I ='t manage this one a bit; it writes all manner of things that I don't intend. J.Z.l T1te meaning of substitute one/ones The substitute OM/ones presupposes some noun that is to function as Head in the nominal group. It is a substitution counter put in to fill the • Head." slot. The meaning is • the noun to fill this slot wiJl he found in the pre- ceding text {occasionally elsewhere) •. In the typical instance the substitute • carries over' only the Head itself; it does not carry over any modifying dements by which this may have been accompanied. So for example in [3 :4] the use of ones as substitute speci£ically excludes the defining Modifier tnAtk of platinum; ones replaces bullets and that is alL Furthermore, however, in place of the original modifying elements the substitute regularly brings with it its own defining
3.2 NbMJNAL SUBSTITUTION 93 Modifier. in this case leaden. The effect is differential; leaden ones is specifi- cally differentiated from (ones that are) made of platinum. h is this differentiation which i.s characteristic of the use of substitutes in generaL A substitute is a carrier of some information which differen- tiates: the instance in which it occurs from the other instance to which it relates by cohesion. In the case of a nominal substitute. this means that it is the carrier of some modifying element which has this differeu.tial function: so. rmes is a •carrier' for kaden which has the function of differentiating the bullets mentioned in this instance from the ones mentioned earlier. those made .if phmnum. It follows therefore that the nominal substitute one/ernes is always accompanied by some modifying element which functions as DEFINING in the particular context. This element is not necessarily the same in its structural function in the nominal group as that which it repudiates; in our example [3: 4], the repudiated element made of platinum is a Qualifier. whereas the one accompanying the substitute, leaden. is a Classifier. (There is a similar example in the sentence just written, where the same thing happens in reverse; here the substitute is accompanied by a Qualifier, namely accompanying the suitllitllte, and what it repudiates is a Classifier, namely repudiated.) Another example: (3 :g] I thought I'd finished with the toughest ass;gnments. They didn't tell me about this one. where the Epithet toughest is repudiated by the Dcictic this. In all such instances the modifying element in the .anaphoric nominal group, namely leaden, accompanying the substiture, and this, is acting as a defining Modifier. We have used the term REPUDIATION, and this concept provides a key to the understanding of substitution (including ellipsis), distinguishing it at the same time rather dearly from reference. The notion of repudiation is explained as follows. In any anaphoric context, something is carried over from a previous instance. What is carried over may be the whole of what there w~ or it may be only a part of it; and if it is only a part of it, then the remainder, that which is not carried ovex:,lus to be RBPUDIA TED. Fur example, in [3: 10] WehavenocoaUires;.onlywood ones. fires ls c.lrried over anaphorically, hut coal i~ repudiated. Semantically this means that. given the set of things designated in the original instance, what is now being designated is in some sense a new mbset. It may be a different subset from that specified previously, as in
SUBSTITUTlON [3: 10); or a subset now specified where none had been specified before, as in [3: rr]. [3: uJ Did you light fires?- On1y- wood ones. It may merely be a new aspect of or angle on what was there before, as in [3: 12] Do you remember that thunderstorm we had the last time we were here ? That was a terrifying one! lt may even be THB SAME subset or aspect, where the sameness is itself unexpected or contrastive. This, interestingly, is the only class of in- stances in which the substitute onefones carries. the tonlc nucleus: [3: 13} Would you like me to change the pictures in your room? - No, I think we'd like to keep the same ONES. Whereas in (3: 10} there is repudiation of an explicit subset,. the class of 'coal fires', in examples such as [3: II-3: 13] what is repudiated is implicit: 'fires other than wood ones'. 'thunde:rstornu other than terrifying ones' (or •thunderstorms in their non-terrifying aspects'), and 'pictures other than the same ones' (ie all those that would result from the process of changing). But what is common to all is that in one way or another there is a redefinition of the • thing' that is: represented by the Head noun, involving some form of repudiation of the definition in the original instance. This: does not necessarily mean that EVERYTHING in the original definition must be repudiated. In [3 :Sa], for example, the presupposed instance is two fat little men~ the presupposing one is: the one marlud •nvM'. Now Tweeclledum is just as: fat and little as Tweedledee is; the only element that i.s: repudiated here is the two. Compare: [3: r.t-] That new cloth-backed Ordnance Survey one-inch tourist map you sold me was ideal - but I gave it away. Have you got another one? where everything is carried over except the that. In instances like this where the Head noun that is: presupposed is accompanied by a string of modifying elements, the context will usuaHy make .i.t dear hQw much is carried over. working backwards,. as it were, from the Head. Exactly the same phenomenon arises in ellipsis. and it will be illustrated further in the next chapter. We said at the beginning of 3.2.1 that the substitute one/ones is a su~ stitution counter filling the Head function in the nominal group, and that
95 it normally carries over only the Head itseJ£ We can now put this more precisely. The use of the substitute always involves some new modifying element that is, therefore, defining: this cne, another one, the biggest one. the Otu' th4t got away and so on. This does not imply that none of the modifying dements can be carried over from the presupposed item. It means merely that there is always some point of contrast; the meaning of the nominal group containing the substitute is never exactly identical with that of the nominal group that is presupposed. This is the essential difference between personal or demonstrative reference and nominal substitutioiL In reference there is a total referential identity between the reference item and that which it presupposes; nothing is to be added to the definition. In substitution there is always some -redefinition. Substitution is used precisely where the reference is not. identical, or there is at least some new specification to be added. This requires a device that is essentially grammatical rather than semantic; the presupposition is at the grammatical levd. The substitute onefones is the marker of a grammatical relation; it presupposes a parti<:ular noun, typically one that is to be fonnd in the preceding text, and is itself merely a kind of counter for which that noun has been exchanged. Since its role is to signal that there is some form of redefinition. it has to be a-ccompanied by some defining Modifier. and can therefore be thought of as a carrier for such defining elements. The process of defining has the effect of re- pudiating whate-ver is not carried over in the presupposition relation: the new definition is contrastive with respect to the original one. For this reason, one can never substitute for a proper name: a proper name is already :fully defined as unique, .and there is no way of adding to or altering the definition. (Oddities like Have you seen Jqhnl - Well, I saw the tall one just now, where there is more than one jolm, are excep- tions not to the use of one but to the general definition of proper names. In this instance John ls being treated hy the respondent as a class name.) 3.2.2 Conditions of use of the nominal substitute As illustration of the use of the nominal substitute rme/ones. Jet us first consider the following forms: [J:Is] (i) a. this one b. the one c. one this new one the new one a new one (ili) this one with wheels the one ""-ith wheels one with wheels
96 Those in row {a) might occur in some such text as Mummy~ wiU you buy me a bus? 1 want ••• AD occur quite freely since each one contllns a modifying element (this, this new, this ... with wheels) to which one is attached as Head~ and which is interpr.eted as defining in function (it was pointed out above that substitute one always requires to be defined). Those in row (b) have the definite article which. as we nw in Chapter 2, does not itself carry the necessary specification; it merely indicates that this specification is available in the environment. In (bii) and (biii) the specification is contained in the Modifier (Epithet new, Qualifier with whuls), to which the refers cataphorically. But (bi) is odd; it has no defining Modifier, and can therefore occur only in a highly restricted context where the meaning is fully specified anapborically~ .and so the sense is •me one you mean'~ as in I know the one, that's the one, these ore the ones. In other contexts only some expanded form such as the one you mean can occur. Those in row (c) have the indefinite article. This is obvious in (cii), since both a and one appear. In (ciii) the indefinite article and the substitute are fused, so that 'a one' is represented simply as qne; it is in fact not im- possible, though lt is still relatively rare. to keep them apart and say a one with wheels. What about (ci)? This: might at £int sight also seem to he a fusion of substitute one with the indefinite article; hut this interpretation will not really stand. For one thing, it is semantically undefined; there is no explicit form to which it can be related, as the ane rdates to the one you HU'an - naturally, since if there was: it wonld have the definite article and not the indefinite. For another thing. it has no p]ural ones. whereas the substitute ont! always: participates in the singnlarjplural system realized as one/ones, eg: I know the ones (you mean). In fact, its plural is some (I want onejl want somt>), and this provides the clue to its interpretation: it is simply the indefinite article in the fonn which it takes as Head of the nominal group. Thus I want one is simply the realization that takes the place of 1 want a; the one is anaphoric. but hy elliplls (not replacement;) of the noun functioning as • Thing~. The relation between substitute one and determin- er (indefinite article) one is discussed further in 3.2.3.3 below. leaving aside (ci), then, on the grounds that it does not contain the substirute one, we can s:ay that in all the other examples under [3 : 15 J the substitute OM is obligatory; as we expressed it earlier, it is a carrier of the specifying dement this, new etc. Even in (ai) it cannot be omitted without changing the meaning. I want this is perfectly grammatical but means 'I want this thing', not 'I want this bus' (or whatever the one in I want this one is substituting for). There are environments. however, in \\o-hich
97 the substitute ont: is optional, giving a choice between substitution and ellipsis: [3 !!6] (i) a. these ones b. which (one) c. hers(?herone) d. Paul's (one) e. each(one) f. the bst(one) g. two(?ones) these new ones which new one her new one Pau1' s new one each new one the last new one two new ones (iii) these(ones) with wheels which one with wheels hers (her one) with wheels Paul's one with wheels each one with wheels the last one with wheels two (ones} with wheels Column (i) is tending to be filled out; one bean not only her one, hU one. and others such as what one. but also. especially in children's speech,. two ones. In column (ii) the substitute is obligatory • since the nominal group cannot normally end with an Epithet. Column (ili) is very variable, and in some instances an alternative of the furm that one of hers with whuls. two of the ones with whals would tend to be preferred. Note that although this does not mean the same as this one (ie- cannot be interpreted as ellipsis). these can mean the same as tJu.sc ones (ie it is ambiguous in contexts where it could be interpreted elliptically): I want these mearu either 'I want these (ones~ eg) buses' or •1 want these things'. As mentioned earlier. the plural of the substitute one is ones. With the exception of [3: I5ci)~ the plural ones could occur in all positions in [3: 15] and [3: 16], showing that these are :indeed all instances of substitution. The nominal substitute can in fact substitute for any count nowt (any noun participating in the number system). either non-human or human; in this it differs from the • pro-noun'~ the only other form of one which has plurnl ones (see J-2-l-4 below). Aside from the doubtful cases of determiner plus substitute? such as my one, twu ones cited above, the one structural environment in the nominal group in which the substitute cannot occur is within a nominal compound: thus we do not normally find examples such as {3: 17] a. Lend me a pen. - I've only got a fountain one. b. let's go and see the bears. The polar ones are over on that rock. c. Is that a tennis racket?- No it's a squash one. d. Are you planting trees here? - I thought of planting some apple ones. This restriction can be stated roughly by reference to tonic accent: a word FOLLOWING the accented word in a nominal expression cannot he sub-
98 ~tituted; and since tonic :accent is itself a re-alization of a compOtmd noun structure, this is equivalent to saying that there can he no substitution within a compound noun. However the restriction also extends, under conditions which are nor dear, to certain instances which on the criterion of tonicity would not be compounds but structures of Classifier plus Thing,eg [3: r8] a. He's an idiot. ~ The village one? h. We sat by a lovely little stream. It was cooJ and dear, like all mountain ones. The tonic structure of vi11age idiot, nwuntain stream, shows that they are Classifier-Noun structures, not compound nouns; and yet it is scarcely possible to substitute the Head noun in such caies. 3.2.3 The wcml one qther than as substitute We have noted that not all occurrences of one are instances of substitutions and it is useful to distinguish the substitute one from the various other words 11ne, the other items which arc forms of the same etymon. These are the personal pronoun one. cardinal numeral one, determiner one (alternative form of the indefinite article) and a fourth MU! which is related to the category of gener.al nouns {see l'i.r below) and which we might refer to as a 'pro-noun', using a hyphen. J.2.J.I PERSONAL PRONOUN one This is the personal form with generalized reference, sometimes called 'generic person\ discussed in Chapter 2 (2.3.1, examples [2-~8]. [2:9]). Another example of it is! [3: I9] One never kno'h>s what might happen. This one has no cohesive force; it is never used anaphorically, but only exophorically, not unlike you and we in their generalized cxophoric sense. It is rather easily distinguishable from the substitute one. since it always occurs 3lone as the sok clement in a nominal group. an environ- ment that is impossible for the substitute, which is always modified (ifp.2 above). J.2.J.2 CARDINAL NUMERAL one This is exemplified in: [3 :20} a.. He made one very good point. h. Ten set out, but only one came back..
99 Again it is clear from inspection that neither of these occurrences of one is a substitute. That in {a) is functioning as (Numerative) Modifier ln the nominal group one very good point, whereas the substitute one functions only as Head; that ln (b) is functioning as Head, but it is unmodified, whereas the substitute is aiways modified. The one in (b) looks at first sight like a substitute because it is clearly anaphoric: there must he some word such as man presupposed for interpretation to be possible. Actually however it is the cardinal numeral one wi:th dlipsis. ln ellipsis, the pre- supposed item, if it is to be made explicit, is ADDED TO the presupposing one: it does not replace it; so ten men set out, but only one- tn4n came back (not but only man came back). In substitution, on the other hand, the pre- supposed item, if it is to be made explicit. RBPLACES the presupposing one: it cannot be added to it; so in [3: 21 J Mummy will you buy me a bus? I want that red one. the non-presupposing form would he that red bus. not that red one bus. In any case once such a nominal group is filled out the distinction becomes obvious, since the muneral always precedes any Epithet that is present. whereas rhe substitute, since it functions as Head, always follows it. Cardinal nwneraf one and substitute one- are quite distinct in meaning. The former contrasts {r) as a Nwneracive, with the other numerals two, thue, etc; (2) as a Deictic, with some, other, both, etc as in: (3: 22] a. The one friend who never let her down was .Enid. b. Can I have those peaches?- You can have one; leave me the other. where again (h) is cohesive by ellipsis, like [3: wb) above. The substitute me enters into no systemic contrasts; on the other hand it may he either singular or plural, whereas the cardinal one is naturally always singular (except in the expression in ones and twcs}. The two mean- mgs are compatible with each other, so we regularly find examples such as [3:2.3] You'vealreadygotoneredone. where the first erne is a numeral and the second a substitute: as long as there is an Epithet or Classifier present, both will occur~ since the numeral n1ust ptteede and the substitute follow. rf there is no Epithet or Classifier~ then the word om- can occur only once; the language has not yet admitted sentences like you've already got one cme, though this will probably occur in the speech of the next generation of children. Meanwhile a form such as you• ve o.lwuiy got ONE might be seen as a fusion of numenl one and
IOO substitute one~ since the meanings are compatible. However~ this inter- pretation will not really stand, for the same reasons as were given for rejecting the interpretation of[J : 15ci] as a fusion of the substitute with the indefinite article. In an example such as you've already got one, there are just two interpretations, the two forms being identical in writing but distinct in speech: either the one is phonologically salient. in which c.ue it is the cardinal numeral, or it is phonologically weak, in which care it is the inddinite article (.see next section). Neither involves substitution, hut both are elliptical. There are other factors diiferentiating the numeral from the substitute. The nwneral accepts submod.ification.. eg: just one? only one, not one, which the substitute does: not; on the other hand the substitute is regularly modified by a Deictic, ~: this one, Yi»lr one, which is rare with the num- eral functioning as Head. Even where the numeral is preceded by a Deictic, there is no ambiguity in speech. for the reason already given: the numeroU is salient, the substitute weak. Hence [3 :24] Have you any envelopes? I need another one. is. ambiguous only in writing; in speech, if one is a numeral. so that the ~filled out • fom1 is another cme envelop~. the tonic will fall on one, whereas if one is a substitute. the filled out form being another envelope, the tonic will fall on another. J.2.J.J INDEI'INTI'E ARTICLE one The normal form of the indefinite article is aftm; etymologically this is a weakened form of the numeral one. The term • article' is somewhat unnecessary, as it suggests that the articles form a separate v.-ord class, whereas both 4 and the are simply members of the more general class of determiner. Within the determiner, a belongs to the non-specific class (including any, elther, no etc). The two major types of determiner. specific and non-specific, embody different number systems. The specific deter- miners distinguish singular/plural~ with 'mass' grouped with the singular, as it is in the noun; so this ho:m, this sugar (singular), these houses (plural). The non-specific determiners distinguish (count) singularfnon-singnlar, with 'mass' being grouped. with the plural; so the form corresponding to a, namely som£, is wed with mass and plural nouns: a house (count singu- lar), S()tne sugat, some houses (non-singular). This som~ is also a phonologi- cally weakened. form. Like many other determiners, the indefinite article can occur elliptically, as Head in the nominal group; for example
3.2 NOMINAL SU:BSTlTUTION IOI [J:zs] a. Are there li.ons in those hills?- Yes, we saw some on the way over. b. I'd like some cof&e. - Then make some. The filled out forms are we saw $0me lions, makE some coffee. What is the equivalent form of the indefinite article when presuppo&ing a count singular noun? [3: 26] a. Are there !ions in those hills? - Yes, we saw one on the way over. b. r d like a cup of coffee.- Then pour yourself one. Here the filled: out forms are a. lion, a cup of a!/Jee (not one lion, one cup of coffie); one is the form taken by the indefinite article when it is functioning as Head of an elliptical nominal group (cf: Have fJne vJ mine!). That this is not the substitute one is shown by the fact that it has tbe non-singular some, and not OMs; cf: Have some of mine! and [3:25a], where we could not say we saw cnes on the way over. Moreover the substitute one does not occur without a Modifier (if[3 :2.1] above). That these forms are elliptical determiners, not substitutes~ is further borne out by the bet that in nega- tive and interrogative they are replaced by any, exactly as the indefinite article is: we JiJn't .see any on tk way owr. (Likewise, the instances where one is :retained in a negative environment are also those where a would be retained in the filled out equivalent. ie where the meaning is specifically singular, <g [3:27] 'I vote the young lady tells us a story." 'Tm afraid I don't know one, • said Alice. -lilled funn I d<m't know a stcry; contrasting with Fm afraid I dun' I ktww any (stories).) This form of the indefinite article is phonologica11y simply a non-salient funn of th.e numeral one. There is therefore an interesting parallelism between the defirute and the indefinite articles in the way they have evolved in Engli.h. Selective form [ salknt J as Modifier or Head (demorutrative) that (numeral) one; some Article [non...Jient] as Modifier [reduced! as Head [weak] the it a/an; [=n] one; some
102 Note the proportionality in the following pairs of examples: (3 :28] a. They need that CHAIR.. They need TRAT. They need the CHAIR.. They NEED it. b. They need one CHAIR. They need ONE. They need a CHAIR. They NEED one. There is thus ambiguity in the written language between one as numeral and one as indefinite article, when functioning as Head of an elliptical nominal group. This is not usually so in speech. because of the difference in salience: the numeral is always salient, and may carry tonic prominence, whereas the article is normally not salient - under certain conditions it can be9 but it can never be tonic. There is typically a contrast between [3 :>!)a and b J : [3 :29] a. fve lost my coat.-// A I/ saw/ one in the/ HALL f yesterday// (=one coat (numeral); non-singular would be ••. f saw J rome ... ) b. fve lost my coat.-// Al J sawoneinthe/ HAll/ yesterday// (=a coat (article); non-singular would be .. . f saw some ... ) As far as the substitute one is concerned~ however. it is distinct from both. by vinue of its occurring only as Head WITH a Modifier present (the one environment that is impossible for numeral or article). Thus in [3: 15] above, all occurrences of one are instances of the substitute except {ci). which is the indefinite article. That the substitute and the indefinite article arc now quite distinct in meaning is shown by the fact that they readi!y eo-occur, as in [3: I5cii] a red one, another one. Where there is no interven- ing element, the normal form of realization is a fusion of the two into a single dement one. as in (ciii) one with wheels; but even here they may be kept discrete, as in I need a tme with a sharp po1'nt. thn-e's .:1 Cnt" lluuln~t seen bifcre. J.2.J.4 'PRo-NOUN• one There is one further meaning of one, in -which it is restricted to human referents; this is not a substitute form. in the sense that it has no cohesive force, but it is not always easy to distinguish it from the substitute one in texts. Examples: {3:30) a. Ifsuchaonehefittogovem,speak.. b. The ones she really loves are her grandparents. Here one means • person' and ones means • people'~ but they are not ana-
IOJ phoric- ther-e is no pres.upposi.tion of an earlier occurrence of the WORD pnson or people or any similar noun. The only other word that functions exactly in this way in .English is thing. These wo:rds one and thing are special items that we might refer to as • pro-nouns •; they are in a sense intermediate between the substitute o~ and the class of general noun discussed in Chapter 6 below (6.I). r~ i.-; this one and thirt~f[ that are found as components in the words something, twthing, anything, everything, some;Jne~ no one etc. Strictly speaking these items one and thing arc members of a class of' pro-furms' which are the equivalent of the interrogative words what, who and so on; the class of 'pro-nouns' thus also includes time, place, way and perhaps reason. Of the two words one and thing, thing corresponds to what and one corresponds to who; hence thing refers to non-human nouns and indefinite nouns. while one refers to definite human nouns. So for example: [3: 3 I J a. What does he need ? The thing he needs (What he needs) is a passport. b. What does he need? The thing he needs is his passport. c. Wha.t doe~ he need? The thing he needs is a lawyer. d. Who doe~ he need? The one he needs is his lawyer. Here the thing c:an be replaced by what; the one cannot, however, be re- placed by who. at least not in modern English. Like the substitute, but unlike an the other forms of one, the pro-noun •ne has plural""''; for example (if[J' 3ob]above) {3:32] Now, my dearest ones; gather round. Since it also functions as Head in the nominal group, and is normally accompanied by some modifying element, lt is easily confused with the substitute (and is generally regarded by grammarians a!! the same item). However, for the purpose of the study of cohesion it is important to keep the two apart, since the substitute is cohesive whereas the pro-noun is not. Moreover there can he ambiguity between them; consider the example {3: 33] The children seemed to enjoy the outing. The one who didn't wasGeorge. Is Georgc one of the children, or is he the teacher? If one is a substitute, it presupp~ child and means 'the child who didn't .. :; if it is a pro- noun, it does not presuppose anything and means • the person who didn't .. .". Given the further fact that the sub;titute is not limited. to
104 SUBST!TU'I'ION human referents, while the pro-form is., it seems desirable to recognize them .as two distinct items. The principal use of the pro-form orw:: in modern English is that exem- plilied in [3:31d] (and [3:33] in its second interpretation): that is, in clauses displaying • theme identification •. * A clause such as the one he needs is his lmvya is related to he neeJs his lawyer in a systematic way; its meaning is that the message consists of two parts, a theme 'his need' and a rheme • his lawyer •, with an equals sign between the two. Compare [3: 34] I know nothing about this scheme. The one you should ask is Dr Rawlinson. meaning • there is someone you should ask, namely ... •. In earlier English it was more widely used in the general sense of' someone •. ie • one of the ones who ... ~, as in Cassius.' [3:35] Hated by one he love~ brav'd by his brother. Such uses have by no means disappeared from the language; but they are more common in written styles than in speech. 3·'·4 S"""""'] of uses of one Here are some further examples to relate the nominal substitute one to the various other items that have been discussed, and to distinguish it from other, non-cohesive fonns of the word one. They are constructed for brevity, which explains (even if it does not excuse) their uninspiring style. [J:36] a. fm fed up with this watch. (I) The thing never wo.b. (2) My old one worked all right. but this one's hopeless. (3) The thing I want now is a solid state microchronometer. (4) P«baps ru get one. b. I like the new manager. {r) The man~s really efficient. (2) The pcevious ones were hopeless, hut this one knows his job. (3) The thing we need now is: some new technicians. (4) Perhaps he'll appoint some. In each example, (1) contains a 'general noun', thing and man. cohesive (~ee Chapter 6); (z) contains two substitutes, one(.r), also cohesive; (3) con- * Sec M. A. K. H:illiday, 'Notes on t£3DSitivity and theme in English'. ]ormt4l D_{Linguistiu J, 1.91"7, ~y Part D. Section6 (pp =J-:z)6).
].2 NOMINAL SUBSTITUTION 105 tairu a ·pro-noun·, thing, not cohesive; (4-) contains an indefinite article, one and sDme, cohesive by ellipsis (see Chapter 4-). The full list of the dements discussed above is given in Table 5· Of those listed, only (r) is a substirute; it is therefore the only one that is properly the subject of the present discussion. (2) is the generic personal pronoun, and is never cohesive. (3) and (4), the cardinal numeral, and (5} and {6), the "indefinite article', may occur in a cohesive context, func- tioning as Head of the nominal group; in that case the form of cohesion is through ellipsis (see Oupter 4). (7). the ·pro-noun', resembles the sub- stitute in having the plural form ones, but it is never cohesive. (8) is the class of' general noun', the members of which regularly enter into cohesive relations; these are treated below, in Chapter 6. In the gre:rt majority of lns.tances. the substitute otu! is anaphoric in orientation. Cataphoric instances are less common; an example would be [3:37] She picked out the loveliest ones of all the 'o"" m the ganlen and gave them to me. where cmes points forward to rtJses. Such instances are however within the structural confines of the sentence, and contribute nothing to cohesion. Finally, we noted earlier that. although substitution is esrentlally a textual rdation, occasional exophoric instances will he found; see the discussion m J.I.I, and example [J :,]. 3.2.5 Nominal substitute same We saw in Chapter 2 that the item same occurs as a cohesive element of the comparative type(2.J.I, examples [2:2t-2]).In such instances. s~ is a rekrence item. not a substitute. There is another cohesive use of same, this time as a nominal substitute, typically accompanied by the. Unlike one~ which presupposes only the noun Head, the same presupposes an entire nominal group including any modifying elements, except such as are explicidy repudiated. Foc example [3: 38] A: I'll have two poached eggs on toast, please. B: tl1 have the same. N~ of course, the same eggs. which would be reference. not substitution. No regular modifying clement may occur with the same; but it is possible to add a reservation to it, and this takes the form of a Qualifier, which is normally inttoduced by "'" and often starts with the word with (addmg a modification) or wiihc>Ut (deleting a mod.mcation. ie repudiating it),
& Table 5: The forms of one, and related items M • M Class Function status nominal substitute Head (always modified) 3 :8-rs I one, ones 3 . .Z, ).2.1 0 • 2 one (they, you, we) personal pronoun Head (never modified) p.J.l 3 :I9 3 ont (two, three .•. ) cardinal numeral Numerative; Modifier or salient ].2.].> J:ZO 4 one, some (both, other) cardinal numeral Deiccic; Modifier or Head salient j.2.J.2 3 :2.2 5 a/on,''"" (any) determiner Dcictic; Modifier redoced j.l.J.J ('indefinite article') 6 one/some (any) determiner Dcictic; Head (never ).2.j.J 3:25-7 ('indefinite article') modified) 7 one,'""' (thing) pro-noun ).2.).4 3: )o-1 8 thing, person, general noun Head (usually with tk) weak(when 6.r 6:x-s atatute, ett anaphoric)
107 eg: the samr but fried, the same (but) without the roast. The presupposed item is almost always non-human. and. it cannot be a proper name. It can, however, be an Attribute: that is, an adjective occurring (in the usual Epithet function) as Head of a nominal group in a clause of ascription: [3 :39] A: Jolm rounded rnther regretfuL B: Yes, Mary sounded the same. Since an adjective is a kind of noun. and rathe,- regretful is a nomina! group, this is still a form of nominal substitution, and so the use of the same in such instances is entirdy to be expected (cf 3·5·2·3 below). There was an earlier we of the same as a pt"onominal reference item, replaceable- as the substitute is not- by him, her, it, or them; eg [3 :40] This is Othello's ancient. as I take it. -The same indeed; a very valiant fellow. This is sometimes imitated. in contemporary usage, especially in the form the very same. Otherwise, this pattern is largely confined today to legal and commercial registers. where the reference is again always non-human and the may be omitted. Note that in this me .same can never carry the tonic: [3:41] We have today dispatched the first consignment of your order. Kindly arrange to accept delivery of same. In Shakespeare's language this pronominal usage with non-human tefer- ence is more general: [3 :42] I am bound to you That you on my behalf would pluck a flower. - In your behalf stiil will I wear the same. J.z.s.:r SAY THB SAME ln the environment of a process in which a 'fact' is involved. the same can often substitute for the fact: for example [3 !43] John thought it was impossible. - Yes, I thought the same, Mot:e often than not one clement in the presupposed clause, usually a nominal, rem.ains outside the domain of the substitution: [3:44] a. We can trust Smith. I wish I could say the same of his pa=. b. Winter ls always so damp.- The same is often true of sum- mer.
.108 In (a) the same substitutes fur 'that we can trust ... • ~ • Smith' being repu- diated by the following of his partner. Similarly in (b) the same substitutes for • is always so damp • ~ with winter being repudiated by summer. One fonn of this usage whiclt is especially common in dialogue is the same opp/ies (to), .•• f!"S for, as m [J:4s]•: His speech didn't say anything new, did it? B: The same applies to most political speeches. All these are devices for making it explicit that the same has the status of a &et. J~2.j.2 DO THE SAME Secondly, the nominal substitute the Slime is often combined with the verb do as a substitute for the process in certain types of dame. An alterna- tive form is d() likew~Y. For example [3 :46] a. They all seatttd shouting. So I did the same. b. My bank manager bought shares in the canal company. Why don't you do likewire? c. That noise really unnerves me. - Yes it does the same to me too. What is being substituted here is the process plus any subsequent element that is not repudiated. This form of substitution is slightly odd, in that what is being substitu- ted is essentially the verbal element in the clause. and yet the structural means is that of nominal and not verbal substitution (for which see the next section). The verb do here is not, in fact. the vecbal substituted" hut the 'general verb~ M, that which occurs in Wlut are fOJI JoingJ Don't Jo that f, r Vi! got nothing ta Jo and SO on; it is the parallel,. in the verb class. to the 'general nouns' thing. per.wn etc (see 6.I). It is distinct from the verbal substitute do in a number of respects. Phonologically, the substitute Jo is weak while the general verb do is salient, Moreover the substitute do substitute> for all verbs except be aod (m British English geoerally) have; whereas the general verb d() is restricted to clause! of ACTION as opposed to SUP"BRVBNHON- essentially, those where the meaning is • someone did something • rather than • something happened to someone'. So we cm have [3 :47} a. I liked the second movement more than I had done the lint. b. That sign means they're: busy - it usually does, anyway.
].2 NOMINAL SUBSTITUTION 109 in both of which do is a verbal substitute; but we could not have. follow- ing (a}~ Yes I did the same, or, following (b). Is that wlw it does? (in both of which Jo is the general verb). because like and mean are not action- type processes. ln some instances there is an equivalent form of nominal substitution for •happening~-type clauses,. with the same (thing) happens, eg h: 48] I lost my way i.n the galleries. - The same thing happened to me. meaning •r also lost my way in the galleries'. Hence the same, although itself a form of nominal substitute. is used as a means of substituting a nominal or other dement in the process as a whole, including the process itsel£ A form such as do the same reflects the general tendency of English to express a process in a nominalized form. hy means of an 'empty' verb pins its object: do a run-through, de a kft turn fo, '•un [it] tlrrough', 'turn ldi:'; if: haw! a fight, give a glance, make a fuss. The presupposing form the same can thus occur as a substitute not only for nominals expressing things, as in [3: 38]. hut also for facts, as in say the same, and for elements that are not strictly nominal at all. Whereas une substitutes just the noun (Head). in the environment of a nominal group having other elements that are contrastive, the same substitutes a nominal group (or something else) in the environment of a clause, so that it is other elements in the clause that provide the con- trastive context. J.2.j.J BB THE SAMB As already pointed out~ the form the same occurs. as Attribute in clauses of ascription, where it may substitute either a noun or an adjective - that is, a nominal group having either noun or adjective as Head, for example (arulifl3:39] above) [3 :49] Charles is now an actor. Given half a chance I would have been the same. Note the potential ambiguity between substitution and reference in such contexts. In [3: so]. if the same is. a substitute the meaning is ~(and) they also taste more bitter than the last ones •. whereas if it is a reference item the meaning is •(but) they bste the same as the last ones did•: [3: 50] These grapefruit smell more bitte' than the last ones we had. -They taste the same.
IIO (In the third possible interpretation. 'they taste like each other·~ which is. improbable here, th£ same is also a reference item, but functioning as Epithet and therefore non-<ohesively; if 2.5.1 above, examples [2: 79] and [z:So).) For the use of S(J in such instances see the next seccion. 3.2.6 Difference between the same and one{s) as nominal substitutes Apart from the type of example illustrated in {3 ; 40 - 3: 42) above, the sanre as substitute is always phonologically salient. It contains an accent. and therefore carries the tonic under typical or ~unmarked • conditions: that is to say, if occurring finally it ls tonic unless rejected fur contrastive reasons. For example, {3 :51] A: I'll have two poached eggs on toast please. B; fU have the SAME. c: r d LlKil to have the same, but ..• n' s utterance has unmarked information focus. In c' s the information structure is marked. with contrastive focus on like. This gives the due to the role of the same in nominal substitution. The substitute one is a grammatical item which contains no accent; i.t is always • given • in meaning, and serves as a peg on which to hang the new infor- mation. In this respect it resembles do (3.3) and so {3-4)~ The substitute the same, however, functions like a lexical item; it can carry the infonnation focus, and typically does so when in final position. The meaning is • the information conveyed by this item in this context is new, but the item itselfhas occurred before'. So for example [3:52] Thencighboursgrow yellow chrysanthemums. a. I could grow RED ones b. I could grow the sAMB c. I could grow some (roo) d. I could grow some of the SAME [n {a) the substitute ones is used so that the Epithet red can carry the UNMARJn!D tonicity: that is. so that the focus of information falls on reJ wnHOUT this becoming contrastive. In (b) the 'yellow' is included in the presupposition, and the substitute the same carries the focus: the informa- tion as a whole is encoded a:; new, with the meaning 'yellow chrysanthe- mums' shown to have been present earlier. (Once again it is s:ubstitution that is appropriate and not reference. They will not be referentiaHy • the !>allle chrysanthemums •; and hence the form I o;JUld grow them would be odd here, with them 'referring, as it were at the lexicogrammatical instead of the semantic level.) In (c) the indefinite article rome is used as
J.Z NOMINAL SUBSTITUTION Ill non-specific Deictic, and the form is elliptical; this focuses information on the I and encodes the whole of the remainder as explicitly given, Finally (d) is like (b), with the addition of the indefinite article as Deictic as in (c). Note that there is no form I could grow red same, since here the 1mmarked tonic would be on same where it should be on red; instead the fonn is I cou.IJ grow the same but reJ, which puts the tonic where it is required while stilllea.ving it unmarked. Essentially the same relation obtains betWeen the same and substitute so as between the same and: one(s}. The patterns discussed in J.z.s.I-3 (say the same. do the same. be the same) are the contexts in which the non- salient substitute alternating with the same would in fact be so .and not one(s). In general so substitutes for a clause, and is dealt with in 3·4 below. However there is no very dear line between nominal and clausal sub- stitution. and these examples are in a way intermediate between the two, In the following the weak form. of the substitute is se: [l: 53l (r) John fdt it was disappointing. (>) John lefi before the end. (3) John sounded regretful (") {•· He SAID so (TOO) 1 b. He said the SAME (;i) {•· MAJ>.Y fdt so (roo) b. Mary fi:lt the SAMB { a. I<lid SO (TOO) b. I did the SAMl! {) {•· He LOOKED so (TOo) 1 b. He looked the SAM:B (ii) {•· MARY sounded 50 (TOO) b. Mary sounded the SAME In type (3) it would be possible to interpret ro following l.,.k, sound, seem. etc as substituting for a clause: Mt~ry sounded as if she was regreiful too. Compare [3: 54] ' ... being so many different sizes in a day is very confusing.' 'lt isn't; said the Caterpillar. 'Well, perhaps you haven~t found it so yet,• said Aliee; ... where so in the Jast line substitutes for VlT}' ccnfosing but could also be filled out as to be very amfosing. On the other hand the presupposed. item need not be of this fo~ as the examples in [3:53} show; and in (3:55] neither the presupposed nor the presupposing item could be expanded into a clause: [3 ~ss] John has become depressed. -Has be ever been so before?
IIZ These are in fact instances in which so is substituting for an Attribute: corresponding to lu! sums intelligent. he semu to be intelligent, he seems as if he U intelligent we have be seems so, he seems to be so~ ht sums as if he is so. If on the other hand the Attribute is represented. by a noun, this must always be a count noun and the weak form is one{..r); so we should have tlaq task .ro (too) in [3: 50) above but I woulJ hav~ bten one (too) in [J :49]. In general~ therefore. although it takes a nominal fonn, tbe si111Je functions as the accented form of the substitute in all types of substitution, clausal and verbal as wdl as nominal: Accented fOrm Non-accented (same salient) Nominal: count noun -(s) th. smne attribute (he) the same Verbal Jo do the same Clausal (reported) (say) the""""' It may also be accompanied by a pro-noun thing (way when substituting for an Attribute)~ as in said the S4tne thing. tDstes tlu! some way; these are constructed like reference items but have come to be used as substitutes in the same way as the iterru same. and so themsdves.. 3. 3 Verbal substitution The verbal substitute in English is do. This operates as Head of a verbal group, in the place that is occupied by the lexical vexb; and its position is always final in the group. Here are two examples from Alice; in both, the substitute is the word that has the form do (not did Ot Jon~t): [3: 56] a. • .. the words did not come the same as they used to do. h. •1 don't know the meaning of half those long words. and, whaes more. I don't believe you do either!' The first do, in (a). substitutes for cotn4!; that in (b) substitutes for know th. m<aning <flutlftlwse 1.mg """"- In [3 =56] the presupposed items are in the -same sentence, and so the substitution is not by itself cohesive. But verbal substitution regularly extends across sentence boundaries~ as in
3·3 VJiRBAL SUBSTITUTION IIJ [3: 57] He never really succeeded in his ambitions. He might have done, one fdt, had It not been for the restlessness of his nature, Here done substitutes for succeeded in his Mnbitions, and so serves to link. the two sentences by anaphora, exacrly in the same way as the nominal substitute one. In the thr-ee succeeding subsections we shall discuss the meaning of verbal substitution, the conditions of use of the verbal substitute, and other uses of the verb do which aie distinct from its use as a substitute (and from which its use as a substitute is derived). J.J.I The meaning of the verhal .substitute do In many wa·ys the verbal substitute do is parallel to the nominal substitute orre. and it is likely that its evolution in Modem English has followed the analogy of one rather closely. There are striking parallels between the structure of the verb.al group and the nominal group in Modem English. although superficially they a.re very different from each other. Like the nominal group. whose structure v.-415 discussed in section 2. r above. the verb.al group has a logical structure consisting of Head and Modifier. and an experiential structure in which the lexical verb expresses the' Thing • .In the case of the nominal group the 'Thing' is typlellly a person, creature~ object, institution or abstraction of some kind, whereas in the verbal group it is typically an action, event or relation; but these a.re simply different subcategories of experiential phenomena, and in any case there is considerable overlap and interchange between the two. ln both nominal group and verbal group, the lexical • Thing~ is sub- stitutable b-y an empty substitution counter that always functions as Head. The substitution form in the nominal group, as we have seen, is ane(s). In the verbal group it is do. with the usual morphological scatter do, does., did, doing. do~. There is a difference between one and do in their potential domains, the extent of the items that they can presuppose. "Whereas one always sub- stitutes for a noun. do may substitute either fur a verb. as in [3: sOOJ, or for a verb plus certain other elements in the dause, as in [3: 56h] and [3:571· At fiot sight it might seem as if tkJ substituted for the whole of what is called the 'predicate' in a Subject-Predicate analysis- the predicator (the verbal gr-oup itself), minus its auxiliaries, together with any comple-
II4 SUBSTn:UTION ments and adjuncts that ar-e present. But any of these may be repudiatetL as the following examples show: [3.:58] a. Does Granny look after you every day?- She can•t do at weekends, because she has to go to her own house. b. Have they r-emoved their furniture? - They have done the desks. but that's all so far. In (a) do substitutes for- look after me but every day is repudiated by ill weekends. [n (b) done substitutes only for removed; their furniture is repudia- ted by the desks. As was pointed out in the discus.'>ion of one, substitution and ellipsis are different manifestations of the same underlying relation, that of presupposition at the lexicogrammatical leveL The use of elliptical forms of the verbal group is very commo~ and there is very little difference in meaning between a verbal group having substitution by do and one having ellipsis (ie substitution by zero). For example [;:59] Inspector (taking back the photograph): You recognize her? Mn Brrling: No. Why should I? Inspector: Of course she might have changed latdy. But I can't believe she could have changed so much. Mn Birling: 1 don't understand you, Inspector. fnspector: You mean you don•t chOQSe to do~ Mrs Birling.* Both elliptical and substitute forms of the verbal group are illustrated here. Why should I? is elliptical. presupposing Why should I rerognize her?; and you Jtm•t choose to do is a substitute form, presupposing you don't choose to understand me. There is very little difference; M could he added in die first and deleted in the second with hardly any change of meaning. Unlike the nomina) group. however. in which under many conditions (eg following an Epithet} ellipsis is generally impossible, and substitution by one is therefore obligatory as the expression of cohesion, in the verbal group there are very few contexts in which the substitute de MUST be used (for these see 3.3.2 below). In general, substitute dt> alternates with zero as a cohesive device, and the meaning is the same in both: the specific process - event, action, relation, etc - that is being referred to must be recovered from the preceding text. If the substitute de is used, its function is to act explicidy as a place-holder. marking out the point at which presupposition is involved. Ir is possib!e to construct examples in * J. B. Priesdey. A .. Iw~C<111s(The Plays ofJ. B. Priestky, VolJ},. Heinemann.
3-3 Vllli.BAL SUBSTITUTION Il$ which its presence appears to resolve what would otherwise be an am- biguity,eg [3:6o] What are you doing here?- We're mycologists, and we're looking for edible mushrooms. - Yes, we are doing too. where the last sentence without the substitute dving would be interpreted 3S 'we are alro mycologists·. Like the nominal substitute cne, the verbal substitute de is typically associated with contrast. It occurs in the context of some other item which contrasts with an dement in the presupposed clause. This is well illustra- ted by instances in which the two clauses,. presupposing and presupposed, are related by comparison, eg a. Mary is doing. b. he should be doing. c. he used to do. [3 :61] John is smoking more now than d. he was doing before. In (a), Mary contrast> with John; in (b), should b. contrasts with is; m {c), used to contrasts with is • .. ing; in (d). was . .. beJtm~ contrasts with is . •• now. Slm.ilar1y do is &equem in the second of two dauses related by before, after, if, when, etc, as in you will finish well before I have done. Since in these cases the two clauses are structurally related, the presence of Jo, while it reinforces this rdationship, is not needed as the cohesive &ctor; but the principle appears dearly. that substitution is a means of representing given information in the environment of new. Exactly the same principle operate:'! where there is no structural relationship between the two clauses, as in examples [3 : 57-6o] a hove~ and here the use of the substitute is precisely what provides the cohesion. Since the substitute is by definition 'given', in that it is a sigruJ that information is to be recovered from elsewhere, .it is phonologically unaccented, or non-prominent. It is usually weak (non-salient) in all positions except when it is the initial, and therefore the only, item in the verbal group; in the latter context it is salient, hue still non-tonic. For example: [3 :62] a. H., anybody fed the cat I- /I SOMBhody / mmt have done /1 b. Did anybody feed the eat I -I/ soMEbody / did 1/ Related to this is the fact that a finite verbal operator preceding substitute do in the verbal group can never be in the reJuced form, since dris would
II6 SUBSTJTUTlON force prominence on to the do: hence forms such as he'll do, he's doing, hl s dt:me cannot occur as substitutes. at least across a sentence boun- dary. However, we saw that with the nominal substitute one there are circumstances under which it is accented, and so can carry the tonic, namely when the • given-ness' of the information it conveys is precisely what is new about it; this is the typical pattern following the same (the same ONES). Likewise with do there is one condition in which it is accented. and as far as the meaning is concerned this is essentially the same condition that we had found with ont', although the form of expression is different: it is when do is followed by so (oo so). The expression do so conveys essentially the same meaning: the action, event or relation in question has been referred to before, but it is precisely here that the new informa- tion lies. So for example [3:63] a. "Yes, I think you'd better leave off.~ said the Gryphon: and Alice was only too glad to do so. b. Just finish off watering those plants. And let me know when you've done so. The expression do sa derives &om pro-verb do (see 3·3·3·3 below) followed by anaphoric so. The do is accented; it is therefore typically salient, with the potentiality of carrying the tonic. The difference in the rhythmic patterns of the two forms of the verh21 substitute, do and do so, can be seen in the following example: [l ~64-} Shall I make an {a. //A you l can do/ NOW 1/ announcement? b. //A you can/ do so/ Now i/ In many irutances either ® or do SQ can occur~ with only a slight difference in meaning: the form with so combines anaphora with promi- nence. so that it has the effect of explicitness, of specifying tha.t it is precisdy the verbal element mentioned earlier that is the point of informa- tion here. But for this very reason there are certain instances where so is obligatory. They are those where .do is REQUIRED to be the point of infor- mation because there is no element of contrast present. .as in [3 :63b) a hove. Elsewhere so is optional; and there are two conditions under which it cannot occur. The first of these is in a comparative clause with than or as. such as [3: 56] and. b :61] above. In fact the form with .w is less frequent in all cases where the presupposing clause is structurally related to the presupposed one; so
J.J VR'R.BAL SUBSTITUTION H7 [3 :65] I want to read this document. You can sign it after tve done so. is likdy to mean ·afier I've read it• (with the tonic on done) rather than 'after l've signed it' (with the tonic on I'vt'); the latter meaning would more probably be expressed as after I have done. The second condition where so cannot occur is if the goal is repudiated, as :in h: 58bj. With the provi:ros mentioned here, the verbal substitute may take either fonn, and the choice between do 31ld du so is often made at the phonological level, on purdy rhythmic grounds: because the one form fits into the rhythm of the sentence more smoothly and effectively than the other. 3.3.2 Ccmlitions of use of the verbal substitute- There is considerable variation in the use of the substitute do in contem- porary English. [t appears to have evolved by analogy with the nominal substitute otU", but to be lagging somewhat behind. one in range of uses. perhaps because ellipsis is .almost always an acceptable alternative. In Shakcspearem Eng1ish, the verbal substitute do was much less clearly distinct fi-om the finite verbal operator do (see 3-3·3·4 below} because of the more general use of the .latter in a positive declarative verbal group. ~: as I do liv~ (Modem English as I liv~). Clear instances of the verbal substitute in Shakespeare almost always have the form do .so. In the follow- ing examples. the do would be a substitute in Modern English; but here it was probably in fact an operator; [3 :66] a. Never a woman in Wmdsor knows more of Anne's mind th=Ido. b. Thou makest a testament as worldlings do. & a very broad generalization, the verbal substitute is used more in speech than in writing. and more in British than American English. Within each of the varieties there are wide dialectal and individual differences. In .British English it can substitute for any verb except be, and [m most dialects) except have in the sense of' possess' ; those verbs substi- tute for themselves: [3 :67] a. t•ve been very remiss about this. -I think we all have been. at times. h. I've had serious doubts about this. - I think: we all have bad. at times. c. I had serious doubts about this. - I think we aJl had. at times.
II8 In (c), Ameri=> and some British speakers would substitute haJ by did; but probably no speakers of English would substitute been by .dotw in (a). In American Engli.sh, on the other hand, db does not generally substitute for verbs of the seem class; in British English it does., provided. the follow- ing Attribute is within the domain of the substitution and is not repudia- ted, [3'68] Paula looks very happy. a. She always used to do, I remember. b. She seems happier now than she did last time we met. Many, p=ibly all, American speak= would find do in (a) and did in (b) both impossible; they would prefer the elliptical form used to in {a) and. the repetition of seemed in (b). There is another factor leading to a considerable differentiation between British and American speakers in their use of the verbal substitute. one which is present in [3 :68aJ and, in fuct, in most of the examples of do that have been cited up to now. This relates to the structure of the verbal group in the presupposing clause.. If this consists- of one word only (simple past or present tense in positive declarative mood), then both American and Briti!h speakers regularly use the substitute; eg: does in [3 '69] Does Jean sing? - No, but Mary does. (Because such examples are easily confused with the finite verbal operator de we have been avoiding them up to now. The difterence will be dis- cussed in 3·3·3·4 below.) If it consists of more than one word. so that the substitute would appear following one or more auxiliaries (in the form. do, d<>ing~ or done}. American speakers prefer the elliptical form in which the lexical verb is not substituted but simply omitted. In [3 :57], for example~ the preferred fmm in American English is he might lu!ve-; and in {3 :6ta--<l] elliptical forms would be expected throughout. and likewise in[J,&]a;mdb]. If the presupposlng verbal group is non-finite, ellipsis is under most circumstances impossible. The rule is that an imperfective non-finite verbal group (those in the • participial' form -in g. eg: going, having gotre) cannot be elllptical; a perfective one (those in the • infinitive' fonn with to, eg: to go, to have gt~ne) can be elliptical only if negative, or if following another verb (eg: wani to go) or cataphoric it. For example, in f3:7o]. (d), (e) and(£) could be elliptical, whereas (a), (h) and (c) could noto
[3: 70] I linally called on him. 3·3 VJ!RBAL SUBSTITUTION II9 a. Having done so [feel better. b. I felt bad at not having done (so) before. c. To do so seemed only courteous. d. Not to (do (so)) would have been di>- courteous. e. I have wanted to (do (so)) for a long time. f. It seemed only courteous to (do so}. In (a)-{c) both American and British English would use the substitute. with American preferring the form Jo sa if there is any choice. In (d)-{f) American English tends to use the elliptical form, or else the substitute with so. whereas British speakers select among all three. at least in (d) and (e). with perhaps some preference for the one with Jq alone. (In (a). (c) and (f) the substitute de cannot occur without SQ, for the reasons given earlier: there is no contrasting element present. In (b). {d) and {e) the contrast is provided by not and by wttiued.) There are thw minor but still interesting differences between different varieties of English revolving around the nature- and potential of verbal substitution. The domain of verbal substitution.. as remarked in 3·3-I above. is the lexical verb together with such other elements in the clause as are not repudiated by some contrasting item. In principle any clement can be repudiated in this way, although certain patterns, particularly those in which a Complement is repudiated, sound a little awkward, and ellipsis. where it is possibl~ seems to fit them better than substitution; this is illusttated by the example. in [3:71) (and if[3:58] above): [3::71] a. Can lions climb trees?- No. but leopards can (do). b. Can liODS kill elephants 1 - No, bu< they can (do) giralf<". c. Have they given the lions. their meat? - No, but they have (done) the cheetahs. d. Can Jions kill with their tails?- No, but they can (do) with their paws. In (a) it is the Subject that is repudiated. in (b) and (c) a Complement \direct object' and 'indi<ect object'), and in (d) an Adjunct. The only element in the clause that cannot be repudiated is the Attribute; [3:72.] is impossible. and cohesion could he achieved het-e only lexically. by repeti- tion of the verb seemed: h; 72] Did the lions seem hungry?- No, but they did restless. There are certain contexts~ however, in which irrespective of its fw1ction
~20 S-UBSTITUTION in the modal structure a particular element in the clause cannot be repudiated. so that it is not possible to presuppose the verb without this ocher element also fal1ing within the domain of the presupposition. In [3:73] the clause cootaining a ves:baf substitute is not acceptable following (i) in each instance. although it is acceptable following {ii) : [3:73]a. [>) She'sneverlivedinEngland. }sh h · F ( .. ) She" . ~ ... L-J e as ne m ranee. u s never sung m .&.06&f:>liUlU. b. (i) Youmustn~tputthemonthetable.} Youcandoonthe (ii) You mustn't cut them on the tab1e. bench. c. (i) The door was shutting. } The windows were d~ (ii) The door was falling to pieces. ing too. tt (i) Wecan"tshutthedoor. } We might do the win- (ii) We can~t smash in the door. dows. The reason is that in {i) there is a strong expectancy binding the repudiated dement to the one that is presupposed by the substitute. For example. live expects a.locative; hence it cannot be substituted without at the same time carrying over any item having a •Location" function that is struc- turally associated with it. This does not apply to sing. which has no such expectancy. Nor does it apply to live in company with other types of Adjunct; there is no difficulty about [3:74] You can~t live on what they would p;iy you. You could do on twice as much, maybe. The same principle lies behind the other examp]es; put also presupposes a 1ocative and so cannot be substituted without entailing the 'Location' Adjunct, whereas cut shows no such restriction. In (c) and (d) there is a coHocational expectancy between door .and shut. so we cannot substitute shut without also presupposing tlocr; notice that it makes no difference whether 'the door is Subject or Complement - the relevant role is that of Medium, which is common to both instances. The restriction does not apply to smash in or fall W pieces, which can be substituted on their own while still allowing the door to be repudiated. A11 these are instances of patterns of expectancy between (i) the Process - action. event, etc - and (ii) a particu1ar role that is related to it: in the structure, such as Location or Med.ium2 or a particular item or class of items that functions i.n that role. eg: door as Medium in relation to the Process shut. An item standing in this relation to the process cannot be excluded from the domain of a verbal suhstitution. lt is the repudiation of other elements in the structure which provides
3.3 VERBAL SUBSTITUTION 121 the contrastive environment within which the substitution takes place. In [3 :7I ], for example. each pair of clauses represents a particular distribution of elements into the presupposed. which are within the domain of the substitution. and the contrasted, which are repudiated from this domain: c. d. Presupposed can climb trees lions can kill they have given meat lions can kill Contrasted liom: leopards dephants: giraffes to lions: to cheetahs. with tails: with paws We have illustrated this by reference to elements in the structure of the clause~ but the same principle operates within the verba1 group: the contrast may occur within the systems associated with the verb itsel( such as tense, polarity and modality. So for example: {3:75] Have you called the doctor?- I haven•t done yet, but I will do. - I think you should do. Here the lexical verb ctlil is presupposed throughout, but there is repudia- tion first of the polarity {negative haven't contrasting with positive have), then of the tense (futuce wiU oontrastiug with past in present have ... -et!). and finally of the modality (modalized should contrasting with non- modalized). The one system that is subject to restriction here is that of voice~ for the reason that substitution is not possible in the passive. Nor- mally therefore there is no change of voice between the presupposed and the presupposing clause; both are active. It is possible however for a passive verbal group to be substituted in the active. for example b: 76] Has she doctor been called by anyone? - I don't know. I haven't done. Maybe someone else has done. There is a tendency in the history of English for active forms of the verb. which evolve first, to he matched by corresponding passive forms after an interval; this has happened consistently with the tense system, and it may be that we are just beginning to see verbal substitution :introduced in the passive. In general. however, while ellipsis occun in the passive in the normal way. subject to exactly the same principles as in the active, substitution does not. · In other respects, a11 the preceding discussion applies to ellipsis as much as it does to substitution. and many speakers wou1d tend to prefer eUiptical foi111S in many of the examples cited. (It is safe to assert. however.
!22 that many people who reject the substitute when their auention is drawn to it actually make frequent use of i.t in their own speech, including in those very contexts in which they claim not to do.} We have stressed all along that ellipsis and substitution are essentially the same rdation, so that it is not surprising to find both as alternative forms of cohesion in broadly the same Tange of contexts. [n the next chapter we shall bring up various other contexts in V~.-hich there is no explicit substitution but only ellipsis. With respect to the use of the substitute, the proviso made in the pre- vious section concerning do and de SQ applies equal1y where the contrast is within the verbal group. Wherever the focus of information is required to bU on the Head of the verbal group- the lexical verb itself. as opposed to :m auxiliary - the substitute takes the form do so. This is sometimes determined by the context, but sometimes appears as an independent choice, as in . ,{•· IIA If didn't {know 1/ WAS doing 1/ h :77] Why do you snule · b. L1A I J didn't 1 know I was J DOING so // where the answer (a) treats the polariry-tense complex {present in past, positive) as the focus of mfurmation, and hence treats smile as simply given, while (b) focuses on smile as precisely the element in which the information resides. There is one other condition which tends to impose prominence on the lexical verb and thus to demand t1o so as the substitute form: this is when the mood of the presupposing clause is other than declarative, ie when it is interrogative or imperative. The reason is that if the verb in an interrogative or imperative clause is anaphoric; the con- trast normally resides in the mood itself, and hence is located within the verbal group: { a, Haven~t you done so? [3:78] Sh:dl! call the d~? b. When will you do so? c. Please do so, as soon as possible. The various conditions: on the use of the verbal substitute, leaving aside variations between different forms, such as de so and do, or zero, and between different dialects or individual speakers, resolve themselves into what are essentially manifestations of the same underlying principle: that of continuiry in the environment of contrast. The continuity, ob- viously, is provided by substitution as a cohesive agency: the rep1acement of the verb by a substitution counter signalling that the relevant item is to be recovered &om elsewhere. But, as in nominal substitution, the signi6-
3·3 VERBAL SUBSTITUTION I2J cance of this continuity lies in the f..tct that its context .is one of non- continuity~ or contrast: some entity or circumstance associated with the process expressed by the verb, or some internal condition of time. mood. polarity or the like. is not as it was in the previous instance. (This is the major distinction between the memings of substitution and. reference as cohesive devices; see Chapter 7 below.) It follows that. if there are certain elements so closely bonded with the Process that they cannot he variecl while the latter is kept constant, they cannot provide a contrasting environmen~ and hence cannot be repudia- ted under conditions of substitution. We referred to these above. It should follow also that the continuity that is being expressed. since it is in the environment of contrast, is not mere -reference back but positive confirmation, a marking of the fact that the lexical verb still holds good This can be seen to be the case. if we consider one further set of examples. [3 :79] a. Smith isn't playing tiddlywinks for his health. He is (doing) for money. h. Were you talking to me? - No. I was (doing) to mysd£ .By any nOC'Iml! interpretation these are wrong. Instead of substitution, some form of reference should have been used: he~ s doing it or hl splaying in (a). I was talking W myHlf in (b). Why? Because the substitute form of (a) means 'Yes he is playing.. but it~s foe money .. ; its information structure is that of What Smith is doing = playing tiJJlywinks. and for money. This applies to both the elliptical form and that with the verbal substitute tking. The required meaning however is • No~ he's playing -only for money'~ with the information structure of What Smith is playing tiddlywinksfor = nwttey. In other words. the process playing tiddlywinks is not part of the information content of the message; it is not marked out for confirmation. but merdy used as a peg on which to hang the informa- tion contained infer nwney. Similarly in {b) the information structure is not What I was !king = talking~ but to myself, which is what substitution implies, but Tht cme I was talking to = myffl/ This demonstrates the principle- on which substitution is- based, and explains the types of limita- tion that there are on its use. Like the nominal substitute qne, the verbal substitute Jo is one of a number of rdat:ed items: lexiul verb. • pro-verb • and so on. A summaty account of these is given in the next section. J~J-3 The wonl do other than as substitute In .addition to functioning as the verbal substitute. the verb tk occurs in
Z24- Modem English as lexical verb, general verb, • pro-verb • and verbal opera- tor. These are all related to each other and form a continuum, or at least a duster, of meanings that shade into one another at the edges.. yielding various indeterminate instances. But the distinctions are significant fur the construction of text. so we will discuss each of these forms briefly in turn (3-3 ·3· I-3-3-3-4)- 3·3·3·1 UCO:CAL VERB Jo This is an ordinary verb of the English language, found in examples such as he has done the job. I have work tc th. let's Jc the tJCCOUnts. Other than in two special meanings, (i) as in do well, th lxully, (ii) as in tlutt will do, wi1l it do?, it will never do to let them know, it is always transitive: it has an inherent Goal. In an active clause this Goal functions as a Complement of the • direct object' type; and since under normal circumstances in English the Goal-Compkment cannot be omitted if the Head verb is expressed (eg we cannot say Have you mended the garage door yet?- Yes, I've mended.), this helps in distinguishing lexical Jc :from substitute do. The tking in [3 :So] must be the verbal substitute. [J: 8oj He ought to be doing his homework. - He rs doing. Ifit had been lexical Jq the form wou1d have been He':s doing it. This would have been recognizable as not being the substitute because of the tonic prominence in doing, the reduced :form of is, and the presence of the non- contrastive Complement it - as we have seen. only contrasting items can occur in the environrnent of the substitute do (except as Subject,. since English normally requires the Subject to be made explicit in indicative clauses whether linked by presupposition or not). Likewise if the answer had been in the passive, eg: it's being done Mw, this could only be lexical do. It :is perhaps worth remarking here on the fact that substitute do can substitute for lexical do in the same way as it can fOr other verbs. Ambiguity may arise where a Complement is present as in [3 :8I] I don't think he likes his new employer much. - No, but he does his job. where does could either be a substitute fur likes ('be likes his job') or lexicaJ do in the expression do the job. Lexical do has in itself no cohesive significance. other than through r<petition (Chapter 6). 3·3·3·2 GBNERAL VERB tJo This is a member of a small class of verbs, equivalent to the class of
3·3 VEltBAL SUBSTITUTION 12.5 general nouns referred to in 3.2 (and discussed in 6.I below). They are lexical items with generalized meanings. This form do occurs in expressions such as they did a donee, meaning simply •they danced', they do lunches • they provide lunches •. it does no harm. Other verbs in the class include make, as in make a mistake 'err', have as in have a bath •bathe", take in take exception to. An example from Alia: {3: Sz] • A little kindness - and putting her hair in papers - would do wonders with her.· 3·3·3·3 PRo-VERI! do Again,. this class corresponds to an equivalent nominal class. that of pro-nouns (3.2.3.4 above). The only members of the class of • pro-verb • are Jo and happen. These stand for any unidentified or unspecified process. do for actions and hP.ppt'n Wr events (or for acti-ons encoded receptively. in some kind of passive form). Their occurrence does not necessarily involve an anaphoric or cataphoric reference; there is nothing cohesive about their use in the following exampJes: [J: 83] a. What wos she doing?- She wasn't doing anything. b. What's happening? - Nothing's happening. c. 'What am I to do r exclai.med A lice, looking about her in great perplexity. However, pro-verb tk is. often used. endophorically. in that it functions as a carrier for anaphoric ite~ especially it and that. The expressions do that, do it in fact function as. reference items.; there are no • reference verbs~ in the language, so we say he did it because we cannot say he itteJ, and hC' dues that because we cannot say he tlkrts. Examples: [3 :84-] :a. 'She~s. tired, poor thing!' said the Red Queen. •smooth her hair - }end her your nightcap - and sing her a soothing lullaby.' 'I haven't got a nightcap with me,' said Alice. as she tried to obey the first direction; ·and I don't know any soothing lullabies.• 'I must do it myself. then/ said the Red Queen. b. Her chin was pressed so dosdy against her foot, that there was hardly room to open her mouth; but she did it at last, .•. c. 'They lived on treacle: said the Dormouse, after thinking a minute or two.
126 'They couldn't have done that, you know.· Alice gently remarked: • they'd have been ill.' 'So they were,' said the Dormouse; • very ill.' d. Whenever the horse stopped (which it did. very often). he fell offin front. The mechanism of cohesion, in such instances, is through the use of the reference items it and that~ but it is really the verbal group as a whole that refers back. so that we could regard do it and do that as compound reference verbs. An occurrence of do thm constitutes :a. single cohesive tie, not two. It is the pro-verb do that occurs in the expressions Jo the same (3.2.5.2) and r1o so (3·3-I). It combines with the pro-noun thing in the expressions Jo scmething, - anything, - nothing~ - 4 ( ••• ) thing and (semanticaiiy the Ull1e dement) what ... tW? Here it often occurs in cataphoric contexts. for example: [3:85]'The first thing I've got to do,' said Alice to henelf ... 'i, to grow to my right size again." The pro-verb do occurs regularly in the passive, for example [3: 86] I told someone to feed the cat. Has it been done? the active equivalent being Has someone J.me it? Here it refers anaphorically to 'the feeding -of the cat'; this is the it with extended reference (if 2.3.3.1 above, [2: t8]} • .and it is perhaps worth pointing out the distinction between a pro-verb do with this type of it and a general verb do with it in a simple pronominal context. In example [3: 87] {3:87) 1 want to make a paper dWn. But it an't be done ln a hurry. the second clause is ambiguous; it is either (i) it r the making of a paper chain'} can't be done (pro-verb) in a huny. or (ii) it(' a paper ch.in') can't be done (gen=l verb, '=de') in a hurry. The distinction is clear in the plum!: [3:88] lwanttom.akesome { (i) Butitcan'tbedoneinahurry. paper chains. (ii) But they can•t be done in a hurry. The pro-verb Join combination with a reference item it. this or that may be anaphoric to any process of the action type. The general verb do is a.naphoric only by lexical cohesion~ in that it stands as a synonym for a set of more specifi<: verbs, as in do sums, Jo an essay~ do the vegetobks, or
343 VBR.BAL SUBSTITUTION I27 combines with them in their nominalized forms, as in de the cooldng, do the writing. and. therefore coheres with -such items if they have occurred in the preceding text. All such anaphoric instcmces of general verb or p~verb do are instances of reference, not of substitution. 3·3·34 VERBAL OPERATOR do The last of the words having the form do is the finite verbal operator or 'auxiliary •. This is in principle totally distinct from all the others, in that it is a purely grammatical element whose function i-s to express. simple present or past tense i.n -specific contexts: when interrogative (do you know?), negative (you don't know) or marked positive (y.ou do know); for example [J :89) Does she sing? - She doesn't sing. - She does sing. This M is always finite. and always occurs as fmr word in the verbal group; it can never represent the lexical element in the process (the .. I'Itmg'). lt would therefore be totally distinct from substitute do were it not for the &et of ellipsi~ fJ : 89] might he rewritten as {3 :90] Does she $ing?- No. she doesn't. - Y ~. she does. where .she det:sn~ l and she d«s are elliptical forms having the operator do as Head. The distinction between this and substitution appears in: [ I D he . , {a. Yes, she d<Jes. 3:91 oess smg. b. No,butMarydoes. [3:9u] is elliptical; doer is the operator and. since it is elliptical,. sing or the substitute do could he added after it: she Joes sing, she does do. [3 :91b] is. a substitute form; sing could not be added after does, but it could .replace it: No, but Mary sings. In speech the two types are more distinct than in writing, because the substitute do is weak whereas the operator do is salient if it is final in theverbal group{k,if eUiptical in the dedarative). An e11iptical interrogative form, such as Does Mary?, is quite unambiguous. because the 'Substitute could never occur in this. form: the interrogative of Mary does, with does as substitute, is dou Mary tk? (consisting of operator + Mary + substitute as Head). The auxiliary do is not itself in any sense a cohesive agent. But the type of elliptical verbal group in which the operator occurs alone is extremely :&equent, and this of course is cohesive by virtue of being ellipticaL This is discussed in 4· 3.
128 3·3·4· Summary of uses of do The set of rdated words do can be illustrated in the following passage: [3:92] What's John doing these days? (t) John's doing a full-time job at the works. (z) That11 do him good. (3) I'm glad he's doing something. (4-) Does he Jike it there? (5) He likes it .more than I would ever do. Here (1) contains thelexical verb do, (z) the general verb~ (3) the pro-verb, which is also present in the original question, (4) the operator and (5) the substitute. Table 6 gives a summary of the items discussed in this section. Of those listed. (1) and (2) are substitutes, and are the subject of this chapter; they are normally cohesive (see below). (3) and (4) are lexical items and are cohesive only in the special context of lexical cohesion; see Chapter 6. (5}. the 'pro-verb', is not in itself cohesive; hut it regularly combines with reference items, ~rticularly it and that, to form what is in effect a verb of reference which is typically anaphoric and cohesive. (6), the verbal operator,likewise has no cohesive force; but it figures prominently in elliptical forms of the verbal group which are thel115elves cohesive by virtue of the ellipsis; see Chapter 4· The substitute do is a! most always anaphoric; it may presuppose an element within the same sentence as itself, so that there is already a structural relation linking the presupposed to the presupposing clauses; but it frequently subrt:itutes for an element in a preceding sentence, and therefore it is, like the nominal substitute, a primary source of cohesion within a text. Only occasionally is it cataphoric, and then only within the sentence • .and so making no contribution to cohesion; an example is [J :93] Since I have done, will you join too? lt occurs exophorically nnder appropriate conditions, for example a warning to someone who has been caught doing something forbidden: [l :94] hhouldo.tdo, if I w.s you. Here the speaker is simulating a textual relation in order to suggest that the action in question is. already under discussion. But its primary function is anaphoric, and it Is a rich source of continuity in everyday linguistic interaction. There is one further type of substitution. that of a clause, which is discussed m the next section (3-4).
Table 6: The fonns of do Function in Class verbal group status r do verbal substitute ).).1-2 2 do so verbal substitute J.J.I w • 3 do lexical verb l·l·J.I [J:8o-I) w < ~ • 4 do (make, take, etc) general verb [3:82] 3·3·P ; do (happen) pro-verb [3:83-6] d 3·3·3·3 = "' " 6 do verbal operator Finiteness: Modifier [3: 89-90] J.J.j.6 d (auxiliary) (Head if elliptical) 0 z -ll
130 3 .4 Clausal substitution There is one further type of substitution in which what is presupposed is not an dcment within the clause but an entire clause. The words used as substitutes are so- and not. (For yes and rro see 3·4·3·I and 4·4·3 below.) 3·4·1· Difference between clausal and llther types of subllitution We pointed out at the beginning of the chapter that, since substitution is a formal rdation. contrasting in this respect with reference which is a semantic one. a substitute typically has the same structural function as that for which :it is substituting. So, for example, one functions. as Head of a nominal group and substitutes for a noun which was Head of a nomi- nal group. ln the same way Jo functions as Head of a verbal group and substitutes for a verb which was Head of a verbal group. In the case of do, however, the substitution may extend over other elements in the dause: any complements or adjuncts that are not repudia- ted fall within the domain of the substitute M. The verb do thus comes close to functioning as a substitute for an entire clause, but for the cule of English grammar which requires the Subject to be made explicit. In an example such as [3:95] The chHdren work very hard in the garden.- They must do. the children falls within what is presupposed in the second sentence as clearly as the other elements do, but it has to be expressed by the personal pronoun they. However. do is not a clausal substitute. This is not because of the require- ment of a Subject, but foe another, mace significant reason: namely that with Jo the contrastive clement which provides the context for the sub- stitution is located within the same clause. It may be within or outside the verbal group. but it is always in the clause itsel£ This was illwtrated in [3: 58] abo-ve. Although other dements may fall within its domain, do is a verbal not a clausal substitute. In clausal substitution the entire clause is presupposed. and the con- trasting element is outside the clause. For example, [3 :96] Is there going to he an earthquake?- It says so. Here the so presupposes the whole of the clause there's going to be an emth- IJUtlhe, and the contrastive environment is provided by the says which is outside it.
r31 There are three environments in which clausal substitution takes place: report. condition and modality. In each of these environments it may take either of two fOrtns. positive oc negative; the positive is expressed by so. the negative by not. We shaH consider each of these in tum. 3.4-1.1 SUBSTITUTION OF lU!POJU'EP CLAUSES Here are three examples &om Alice: [3:97] a. • ... if you've seen them so often. of course you know what they're like'. 'I believe so,' Alice replied thoughtfully. h. • How am I to get in?' asked Alice again. in a louder tone. •Are you to get in at all?' said the Footman. 'That•s the first question, you know. • It was, no doubt: only Alice did not like to he told so. c. 'The trial cannot proceed,' said the King in a very grave voice~ • until all the jurymen are back in their proper places - all.' he repeated with great emphasis. looking hard at Alice as he said so. In (a), so substitutes fo< (that) I know what they're like; in (h), for that was the first questitm; and in (c) for' all'. As example (c) shows, the presupposed dement nuy be in the quoted form ("direct speech). The reported clause that is substituted by .ro or not is always declarative, whatever the mood of the presupposed clause. There is no substitution fur interrogative or imperative (indirect questions or comma.nds), and therefore the clause substitutes do not occur following verbs such. as Wl>fflleT, orkror ask. The essential dittinction to be nude here is that between repom and facts. This is a complex distinction. hut it is fundamental to language and is rdlected in the linguistic system in very nuny ways. Broadly speaking, bets and reports are those elements in a linguistic structure which represent not the phenomena of experience themselves - persons, objects, actions, events, etc - like 'children •, • throw • and • stones' in the chiltlren were throwing stonu. but such phenomena already encoded in language, fOr example 'the fact that the children were throwing stonei' as in (the f4Ut) tluzt the chiltlren were throwing stones displeaseJ their parents. These encodings then participate in linguistic structures in the normal way, as this example illustrates, although there are dear restrictions on the types of clause into which they can enter. What matters here is that these encodings are of two kinds: facts,
132 and reports. The two are rather different, though they are not always easy to tell apart. Facts are phenomena that are encoded at the semantic level, as meanings; reports are phenomena encoded at the lexicogramma- ricaJ Jev.d, as wordings. Report corresponds more or less to the concept of • speech • in 'direct speech' and 'indirect speech'. Here are some examples of both; those in column (i) are reports. those in column (ii) are facts: [1:g8] (i) Report a. Mary said: •Jolm'slate.' h. Mary'sas~rtion: 'John is late.' c. Mary said that John was late. d. Mary's assertion that John was late. e. f. g. Mary was afraid that John was late. h. Mary's fear that:Jobn was h1te. · (ii) Fact Mary resented that Jolm was late. Mary's resentment that John was late. Ma:ry resented John's lateness. It was that John was late that Mary resented. Mary was angry that John was late. J· That John was late angered Mary. 1t will he seen that, although the typical form of expression for facts and reports is the same, as 11lustrated in (c), (d) and (g), there are other realiza- tions which are restricted to one or the other. The restrictions are not as totally dearcut as they have been made to appear here, because other factors are involved as weU; but they are valid in generaL and they follow from the general di:ttinction between fact as meaning and report as wording. 'Ibis last formulari<m should not be taken to imply that a report alwa~ follows the exact wording of what was said, or that there necessarily was an act of speaking correspon<ling to it. Reports are associated with think- ing as weU as wi.th saying. It merely means that facts are semantic struc- tures while reports are lexicogrammatical str1.1Cttlres. And this enables us to predict, '-'.'hat is actually the case. that reports can be substituted whereas facts cannot - since. as we have seen. substitution is a lexicogrammatical relation. Hence, corresponding to (c) and (g). we can have 1\lfary said so, Mary was afraid se in colunm (i); but we cannot have Mary resented ;W or Mary was angry so in column. (ii). (The fact that so could not appear in (b) and
133 (h) is a purely structural limitation; we cannot say Mary'sftar so because here so would be a Qualifier within the nominal group, which is not a possible strurtural function for it.) The pattern appears dearly with words that are used to introduce both facts and reports, eg: regret, which means either •be sorry about the fact that' or 'be sorry to say that'; the substitute can be used only in the latter sense. Facts can be presupposed by reference. but not by substitution: {3:99] a. They've failed. then?- I regret so. b. They've failed. - I regret it. - Everyone regrets it. The negative form of the clausal substitute is not, as in (3 :100) Has everyone gone home?- I hope not. However with some verbs: negation tends to be transferred into the reporting clause so that, for example. the normal pattern of think plus negative substitute is don't think so rather than think not; if also I don't believe/suppose/imagine so. The word not can be interpreted .as the 'port- manteau' realization of the substitute and negative polarity. All such expressions:, positive and negative, are particularly frequent in first person singular, where their meaning comes very dose to that of expressions of modality (if 3·4·1-3 below). Another example: [3: 101] 'Of coune you agree to have a battle?' Tweedledum said in a calmer tone. •r suppose so.· the other sulkily replied. as he crawled out of the umbrella. There is some restriction on the use of the substitute in the context of expressions of certainty; we say I' m afraid so but not I" m sure so, you think not or you don't think so but not you know not or you JM't know so. The same restriction turns up with modality, though only in the positive; we say perhaps Stl hut not certainly so, although here certainly not is regular and frequent. This is perhaps correlated with the distinction between facts and reports: .a report that has certainty ascribed to it strongly resembles a fact - unless the certainty lies in its negation. But the pattern is by no means a consistent one. One type of report in which substitution is especia11y frequent is the impersonal type. eg: they say sojtwt, it says so/not, it seems/appears so/mt; including. rather more restricted.ly. those in the passive form: it was reported so, it is .said not. For example, h : I02 J Ought we to declare our winnings? - It says not.
IJ4 There is a possibility of overlap between this structure and that in which so is substituting for an Attribute in a clause of ascription; for exantpl~ in [3: 103] Is this mango ripe? -It seems so. the answer is strictly speaking ambiguous: it may be 'this mango seems ripe·. with personal reference item it and so as nominal substitute, or 'it seems that this m,ango .is ripe' with impersonal (non-anaphoric) it and .ro as clausal substitute. The distinction becomes clear in the plural: Are these mangoes ripe? - (i) They sum so. (ii) It seems so. The difference in meaning is slight, hut it is easily perceived. Note that the negative is not in both instances, (i} They seem net. (ii} It seems not; showing that both are in fact substitution forms. Finally. so as a report substitute occurs in initial position in expressions such as se it seems, so he .said, .so 1 believe, so we wer~ led to understatul. This has the effect of making the .se thematic in the clause. Since negation when not combined with other meanings is rarely thematic in modem English. there is no equivalent negative form. 3.4..1.2 SUBSTITUTION OF CONDffiONAL CLAUSES A second context for clausal substitution is clut of conditional structure. Conditional clauses are frequently substituted by so and rwt. especially following if but also in other forms such as assuming so. suppose rwt: {3: 104] a. Everyone seems to think he's guilty. If so, no doubt he'll offer to resign.. h. We should. recognize the place when we come to it.- Yes, but supposing not: then what do we do? Hcre so in {a) substitutes for he is guilty, not in (b) for we Jcn't recognize the place when we come to it. 3-4.!.3 SUBSTITUTION OF MODAI.IZBD CLAUSES Finally. so and rwt occur as substitutes fur clauses expressing modality, eg [J: 105] a. 'Oh, I beg your pardon!' cried Alice hastily, afraid that she had hurt the poor animal's feelings. 'I quite forgot you didn't like cats.' • Not like cats ! · cried the Mouse, in a shrill, passionate voice. 'Would you like cats if you were me?' 'Well, perhaps not.," said Alice in a roothing tone: ...
I)S b. • May I give you a slice?' she said, taking up the knife and fork, and Joolcing from one Queen to the other. 'Certainly not,' the Red Queen said, very decidedly: 'it isn't etiquette to cut anyone you"ve been introduced to. Remove the joint !' Modality is the speaker's assessment of the probabilities inherent in the situation, as in (a)~ oc, in a derived sense, of the rights and duties, as in (b). These may he expressed either by modal forms of the verb (witl. would, can. ccu/J, m.xy, might, must, should, is t<1~ and ought to), or by modal adverbs such as ~rhaps, possibly, probably, certainly, surely; the latter are frequently followed by a clausal substitute, with the proviso already noted, that those expressing certainty do not accept substitution in the positive, though they do in the negative. 3·4·2 Similarity ammg the types of dausal substitution We have distinguished. the three types or contexts of clausal substitution: report. condition, and modality. It is important to emphasize, however. what they have in common. To start from mocblity: there is considerable similarity in meaning between a modalized clause, on the one hand. and a reported clause dependent on .a 6rst person singular verb of cognition on the other; for example between probably he's right and I S11ppou he•s right- and. hence between probably so and I suppose so. But I suppose is merely a special instance of • someone supposes •; and supposing is merely one way of • cognizing' a report, among a set of possible ways including thinking, assuming, believing, knowing and so on. The unmarked context for a report. however. is one not of cognizing but of verbalizing: not of think- ing, but of saying. Hence there is a semantic continuity, a 'dine', all the way from proftably he's right, through I think he's right and they think he's right to (impersonal) they my he's right and Mary says he's right. All of these can be substituted by so and, in the negative. by not. Looked at from another angle. however, a conditional clause is also semantically odated both to a reported one and to a modalized. one. The form if hi's right means 'let us suppose he·s right; then .. :; the condition may be expressed by non-finite (dependent) forms of verbs of cognition, such as supposing, assuming (and in many languages by verbs of saying, equivalent to 'let it be said that. _ . :then .. .'). Likemse, we can interpret
I36 if he's right as a modality. similar: ro 'possibly he• s right; in that case ... •; and again there is a modalized form for the eA-pression of a conditional: should Ire be right •••• All three types have the property of being at one remove from (state- ments of} reality; they are hypotheticaL Modalizing, reporting and conditionali:zing are all ways of assigning dependent status to the clause in question. This is reflected in the structure; reported and conditional clauses are bothHYPOTACTIC but not • embedded' (ienot RANKSHI PTED; it is this that is the rdevant concept. since • embedding • has not been dearly distinguished from hypotaxls in much recent grammatical analy- sis). That is to say, such a clause is DEPENDENT ON another clause but not structurally integrated into it~ it is not A CONSTITUENT OF it. Since modality is normally expressed within the clause, by a modal Adjunct such as possibly, or by a modal operator in the verbal group such as tm~y, there is no hypotaxis involved; but where modality is expressed by a separate clause. then the modalized clause is likewise hypotacrically related to it, as in it may be that he's right. This then is the general environment for dausal substitution. It occurs in the context of hypotaxis; a clause that is hypotactically rda.ted to another clause may be substituted by so or twt. Semantically this hypot.ac- tic structure is the expression of dependent or hypothetical status, in the form of report, condition or modality; and the possibility of substitution therefore also extends to the other realization of this relation, namely a modalized clause in which the modality is expressed simply by insertion of a modal Adjunct. As with nominal and verbal substitution. the key concept is one of continuity in the environment of contrast. It is nOt possible to substitute a clause which is fUnctioning independently. just because it is bcing repeated; in such instances it must be presupposed by reference, typically by it, this or that. Substitution is used in order to display the clause as a repetition but in a contrastive context~ one in which it is dependent on something else- a report, a condition, an opinion. As alway~ what we are calling • contrast' is not necessarily a negation of the context that was there before; there may have been no such context, and even if there was. the presupposing context may be si.mply a reaffirmation of it. But there is always some redefinition of the environment of the presupposed dause; the speaker or writer is encoding the clause as itself recoverable but in a context which is non-recoverable. This is the underlying meaning of clausal substitution, and it relates it clearly to substitution in the other contexts.. nominal and verbal
T37 3·4·3· Some rdated patterns There are various patterns either related to or in some way resembUng clausal substitution wbi.ch may be brought together for a brief mention here. These fall under two he;zdings.: forms of response, and othcr use:s of se and noJ. 3 .4. 3. I RESPONSE FORMS The following examples illustrate fOrms of response which. could be interpreted as subnftution: (i) [J:to6J a. Hens lay eggs. So they do! b. Hens don•t fiy. So they don't! So do turkeys. Nor/Neither do turkeys. Those in column (ii:) are responses which add a new Subject; the mean- ing is ·and + (Subject) + do so". Since they have alternative forms tmkeys do {.so) tco, tutkeys don't (do se) eitker. they can reasonably be inter- preted as forms of verbal substitution of the do so type (J.J.~ above), with the additional meaning of "and. too'. This meaning is always present in the negative form nor • and not'~ in this structure it is present also as a component in the so, by virtue of its initial position (if the di5Cussion of 4 and' in Chapter 5 below). The examples in column {i) are exdamarory responses, acknowledging new infOrmation arul expressing agreement with it; 'now that you men- tion it, I see you're right'. There is no meaning of • and' here, but some speakers have an alterna- tive form of the negative, namdy twrfneither they do!~ and what may he a subset of the same speakers have tCQ in the positive~ they do too!, (more used in contexts of contradiction) - perhaps this pattern has evolved through influence from the column (ii} forms. The column (i) expression as a whole is undoubtedly cohesive; hut it seems that the cohesio~ here is rather a matter of ellipsis, and that the so is being used in the non-cohesive sense of'true' (see next section). Note that there is a superficially identical structure in which so is a reference ltem, meaning •like that', 'as pre- viously stated', eg (if the last sentence of example lJ :84c) above): [3: 107] It c:an't have helped vcry much. all that shouting.- So it didn't. It only made things worse. The other response forms that need to be mentioned are yes and tw. These could be thought of as clause substitutes; but they are really more
138 readily interpretable a-s elliptical forms. (See Chapter 4. section 4-4-3. for discussion of yes and no.) They express: just the polarity option in the clause. positive or negative. leaving the remainder to be presupposed. It is important to make this dear: what is expressed by yes and no is the polarity of the presupposing clause, irrespective of the polarity of what it pre- supposes - they do not? as their dictionary equivalents in some languages do, express agreement or disagreement with what has gone before. Consequently the response yr:s means • I am tired' no matter what the po1arity and mood of the presupposed clause: [3:108] a. You•retired. b. Are you tired? c. You're not tired? Yes. {I am tired.) d. Aren't you tired? etc The substitutes so and not are exactly paralld to yes and no in this respect; so for example [3: 109] a. Is he going to pass the exam?} {I hope so. (I hope he is.) b. Isn't he going to pass the rm afraid not. (I· m afraid exam? he isn't.) 3·4-3·2. OTHER USES OF so AND :tW! In Chapter 2 we discussed the use of so as a reference item, meaning 'like this' or "to this extent', There is no such thing as negative reference, so the form net does not appear under this heading. In Chapter 5 we shall deal with so as a conjunction~ meaning • conse- quently'. Here too there is no rdated negative form. In this chapter v.-e have treated. so as a substitute, in nominal, verbal and clausa1 sub§titution, in which it stands for the whole or part of another (typically a preceding) clause. Clausal substitution is the only context in which s11 has a corresponding negative, namdy not. It is also the only context in Vi.>hich nat is a cohesive element; elsewhere it is simply the expression of negative polarity. In all these instances, so is cohesive. There remains. one further use of the word, in which it is not cohesive, but simply has the meaning 'true'. (This has already been mentioned in the last section, as the interpretation of so in So they do!) It is this meaning of so that is found in the expressions that is so, this being sa, is th.tt .w?, and so OIL Here the meaning is 'that is true', 'that is the case'. That this is not the substitute so is shown among other things
3.4 CLAUSAL 'StTBST1TUT10N 139 hy the negative; the neg~tive of that is so is never that £.1 twt, but only that is not so~ that isn•t so (in substitution both forms would be expected to occur}~ The cohesiveness of expressions of this kind derives not from the so hut from the anaphoric reference item that or it. In this sense so almost always follows the verb be. It might be possible to interpret it sanu so in this way~ but here there is the corresponding negative #stems not as alternative to it doem""t sum se. Moreover the StJ in it sums ro is typically non-tonic, which suggests the unaccented, substitute form; whereas the so following the verb be is typically tonic, eg: that is Sb-. it's not so. The difference between the two is :shown by the fact that. by itself; it tl so is not a possible response; [3: noJ is unacceptabk: f3: I re] Everyone's leaving.- It is so. whereas -it sums so would be quite acceptable. Perhaps in a sequence such .. [3: rn] Everyone's leaving.- It seems so.- [t is so! where the context sets the tonic on is and not on .so, we have a true substi- tute so following be; but this is clearly a special case, and can be felt to be somewhat odd. 3·4·4 Summary of uses of so The m~ of so are summa.rized in Table ;. Of the items listed, {I-6) are all cohesive; only (7) is not. {r) and (2) are reference items (Chapter 2, sections: 2.5.1 and 2.5.2}. (3) refers to the only use of so as a nominal substitute, as Attribute in ascriptive clauses such as they seem so. In (4). so is the verbal substitute do so, discussed in 3.3; for the form so de I, see 3-4·3-I. (s) is so as the clausal substitute, in contexts of report, condition :md modality; this was treated in 3.4-2. In (6) so is a conjunction (Chapter 5. especially 5.6). In (7). so has the meaning of' true'; it is not cohesive. and so not di$CUSSed in detail, hut the forms it is so and so they do! were mentioned briefly in sections 3-4-3-I and 3·4·3·2· As a clausal substitute, so is almost always anaphoric, exactly as are all substitutes. Like the others. it may presuppose an element to which it is already structurally related; but since it itself substitutes for a clause.. the only conditions under which this can occur are those of structural rela- tions between clauses, paratactic as in (3: r rza]~ or hypotacticasin[J: rrzb]: [3: :nz] a. He may come, but he didn't say so. b. He'll come if he ~id so.
t Table 7: The forms of so ~ • Cohesive type ci .. , Function otatus " " I SO reference Submodifier in 2.5.1-2 [>:87] 0 • nominal group zso reference Adjunct 2.j.I-2 [z:78a]; [3:8o~<]: [3'107] 3 so substitution adverb ).2.6 (3: 54); (J:IOJ) (nominal) 4 so (in do"') substitution adverb (part of) Head of 3·3; ].4.).1 [J:77-1I]; (verbal, verbal group [3 !Io6(ii)] part of) 5 so substitution adverb Adjunct (clawal) saJient 3·4·1-2 [3 :9s-sos] 6so conjunction conjWlct:ive Adjunct s.6 fs:4J..O) 7'0 - 3·4·3·2 [3 :Io6(i)]
3 ·4 CLAUSAL SUBSTITUTION 141 where so substitutes for he would come in each case. (Note that in (a) may is the realintion of • possibly + future • o the substitute presupposes the 'future' but not the modality.) Such instances derive their cohesion internally from the structure. Even more frequently than the other substi- tutes. however. w presupposes across the sentence boundary. and hence functions as the primary me:ms of textual cohesion. Cataphoric instances are infrequent but by no means impossible: [3:n3] Ifhesaidso.he1lcome. But it is difficult to construct exophoric examples, because of the ~rticu lar nature of the contrastive contexts - report, condition and modality - in which clausal substitution occurs. As was mentioned at the beginning of this chapter. substitution is fundamentally a textual relation; the primary meaning is anaphoric. and a substitute is used exophorically only when the speaker wants to simulate the textual rdation in order to create an effect of something having already been mentioned. This rarely happens where the presupposition extends over the meaning of an entire clause. Substitution forms are swnmarized in Table 8. Table 8: Summary of substitution furms Non-prominent Prominent (given) (new) Thing (count noun) one(s) the SAME Nominal Process (nominalized) he the SAMI! fact say Verbal Process ( + . . . ) DOSO Clausal (p); positive report, condition, negative not NOT modality