Upload
univ-lille
View
1
Download
0
Embed Size (px)
Citation preview
UNCORRECTED
PROOF
Chapter 8 1
What’s Pragmatics Doing Outside 2
Constructions? 3
Bert Cappelle 4
Abstract This chapter argues against a view according to which pragmatics, as 5
opposed to semantics, is completely outside grammar. It suggests that, on the 6
contrary, speakers strongly associate various pragmatic aspects of information with 7
constructions. I here give an overview of a wide range of pragmatic phenomena as 8
they have been dealt with in Construction Grammar, a linguistic framework which, 9
as a matter of principle, accommodates pragmatic information in the description of 10
stored form-function units. Such information includes Gricean maxims, information 11
structure, illocutionary force and larger discourse structure. However, Construction 12
Grammarians have been rather vague on what kind of (presumably) pragmatic 13
data should and should not be included in a construction and whether or not, 14
within a given construction, pragmatics and semantics constitute separate layers of 15
information. I demonstrate a heuristic based on cross-linguistic or intra-linguistic 16
comparison of functionally similar constructions (e.g. Can you : : : ? and Are you 17
able to : : : ?) to decide whether we should explicitly specify ‘short-circuited’ usage 18
information (e.g. the request use of Can you : : : ?) that could in principle be obtained 19
purely on the basis of sound reasoning. I also propose that semantics and pragmatics 20
should be treated as distinct levels of functional information in constructions. 21
Keywords Construction Grammar • Gricean maxims • Speech acts 22
• Information structure • Short-circuited interpretation 23
8.1 Introduction 24
Here is a simplistic view of constructions, one which I will not attribute to any 25
particular linguist. In fact, I am not even certain that anyone could be cited to hold 26
this view explicitly. Yet, I do believe it is a conception which is implicit in much 27
work in linguistics, and that spelling it out is therefore warranted. 28
B. Cappelle (�)Université Lille 3, CNRS, UMR 8163 – Savoirs Textes Langage (STL), F-59000 Lille, FranceAQ1e-mail: [email protected]
© Springer International Publishing Switzerland 2017I. Depraetere, R. Salkie (eds.), Semantics and Pragmatics: Drawing a Line,Logic, Argumentation & Reasoning 11, DOI 10.1007/978-3-319-32247-6_8
115
UNCORRECTED
PROOF
116 B. Cappelle
Constructions are units linking form with meaning. The meaning of a construction can 29
be called its semantics. While one may more broadly speak of a construction’s ‘function’ 30
instead of its meaning, constructional semantics is to be distinguished from pragmatics. 31
This latter notion necessarily falls outside of a construction, given that it deals with how the 32
context in which the construction happens to be used may also contribute to its meaning, 33
perhaps better called its ‘interpretation’. 34
This view is problematic, because some constructions have pragmatic content built 35
into them. In this chapter, I will therefore attempt to provide a more satisfactory 36
view of constructions, one which does justice to several ways in which pragmatic, 37
context-dependent effects on the interpretation of a construction can be seen as part 38
of what we know about that construction. 39
So, how can pragmatic meaning be represented in constructions, if at all? The 40
approach taken here to answer this question is qualitative and theoretical. That is, 41
the present chapter does not focus on one particular empirical phenomenon but 42
surveys a range of pragmatic areas (Gricean maxims, information structure, etc.), 43
with the aim of examining if and how different pragmatic tiers of information could 44
be integrated into a description of constructions. In particular, I will examine how 45
pragmatics can be handled by Construction Grammar, a (depending on the variety 46
used) more or less explicit theory of form-function linking. A major question to be 47
resolved is how much pragmatic information speakers have to store in their long- 48
term memory and how much they can just figure out. Interpretations which are in 49
principle computable may nonetheless be ‘short-circuited’. A familiar example is 50
the request interpretation of Can you X? But how can we know whether language 51
users, when hearing such a pattern, make use of a stored interpretation rather 52
than compute it anew on each encounter? We will see that there are often some 53
simple cross- and intralinguistic diagnostics that we can apply as ‘ordinary’ linguists 54
to circumvent psycholinguistic or neurolinguistic experiments designed to probe 55
speakers’ memorized pragmatic knowledge. 56
One construction which is clearly equipped with built-in pragmatics is discussed 57
in a much-quoted paper by Kay and Fillmore (1999) and illustrated by the sentences 58
in (1): 59
60
(1) a. What is this scratch doing on the table?b. What do you think your name is doing in my book?c. I wonder what the salesman will say this house is doing without a kitchen.
(Fillmore and Kay 1999: 3)
61
Kay and Fillmore call the construction used here the What’s X doing Y? (WXDY) 62
construction, named after its special formal features, which include the obligatory 63
use of what and the progressive form of do. This construction doesn’t just have a 64
recognizable form but also comes with a special interpretation: the speaker uses it 65
to make it clear that she judges the situation as being out of the ordinary, to say the 66
least. We immediately understand that the speaker, pointing at a scratch on the table, 67
wishes to express her displeasure at seeing it there: there wasn’t a scratch before, so 68
how come there is one now? Likewise, in the second example, we have no trouble 69
grasping the speaker’s general message that seeing the hearer’s name in her own 70
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 117
book is not what she had expected and that she demands an explanation. And it is 71
the utterer of the third sentence – the actual speaker using this construction, not the 72
salesman – who indicates that a house without a kitchen is an anomaly, and would 73
like to hear the salesman explain this unusual situation. 74
In the old joke about the diner who asks, “What’s this fly doing in my soup?”, 75
the waiter who replies with “Madam, I believe that’s the backstroke” points at a 76
potential ambiguity in the construction. The waiter answers a direct question about 77
the fly’s precise kind of activity. Yet, what the fly is actually doing isn’t something 78
the diner is interested in, as she just wants to indicate her indignation at there being 79
a fly in her soup in the first place. As Kay and Fillmore point out, this special 80
interpretation cannot be the result of a conversational implicature. That is, it is not 81
as if the hearer first considers the speaker’s literal question, then deems this question 82
as being beside the point and finally gets the speaker’s drift, which is that she finds 83
there to be something surprising and incongruous about the scene. The reason why 84
this cannot be what’s going on is that many WXDY sentences, including those in 85
(1a-c), lack such a literal reading: “A scratch, a name, a house, and so on cannot 86
be literally said to be ‘doing’ anything” (Kay and Fillmore 1999: 5). The special 87
interpretational effect of WXDY sentences must instead be analysed as a property 88
of the construction itself. This is also clear from the observation that a WXDY 89
sentence such as What are your children doing playing in my garden? is necessarily 90
an utterance by which the speaker expresses disapproval, while a closely resembling 91
non-WXDY sentence such as Look what your children are doing in my garden may 92
or may not be so interpreted, as it could also easily be followed by How adorable 93
(Kay and Fillmore 1999: 5). 94
Literally intended questions may nonetheless have played a role in how the 95
WXDY construction developed its now-standard special sense. Kay and Fillmore 96
(1999) put it thus: “While the WXDY construction may have had its origin in 97
conversational implicature – through situations in which an individual A is clearly 98
up to no good and B asks what A is doing – the semantics of incongruity 99
is now CONVENTIONALLY associated with the special morphosyntax of WXDY 100
constructs” (Kay and Fillmore 1999: 5; boldface mine). Note that Kay and Fillmore 101
speak of the incongruity judgement in terms of a semantic aspect of the construction, 102
not a pragmatic one. Referring to this passage in Kay and Fillmore’s article, Bybee 103
(2013) further comments: 104
The conventionalization of implicature (or from the hearer’s point of view, inference) is also 105
well-known from grammaticalization research [references omitted – B. C.]. It is thought that 106
the frequent co-occurrence of an inference with a particular construction can lead to that 107
inference being taken as part of the meaning of the construction. The originally inferential 108
meaning can even replace the earlier meaning. (Bybee 2013: 56) 109
Observe again how, in this view, a once-pragmatic interpretation can become an 110
aspect of the semantics of a construction or even supersede the erstwhile semantic 111
information. 112
This leads to two simple but fundamental questions. First, can we ever claim 113
that something pragmatic is part of a construction and is thus an aspect of stored 114
How adorable
UNCORRECTED
PROOF
118 B. Cappelle
language knowledge, rather than being a context-dependent interpretation the hearer 115
computes on the fly? Second, supposing that pragmatic information can indeed 116
be incorporated in constructions, does it inevitably become something purely 117
semantic – an inherent aspect of meaning – or can we still distinguish between 118
semantic and pragmatic aspects of a construction? 119
The first question is meant to resolve an apparent conflict between the notion 120
of pragmatics and the notion of constructions. Pragmatics is about how the hearer 121
makes sense of an utterance by applying general cognitive principles to linguistic 122
elements as they occur in a specific context of use. Constructions, in their turn, are 123
viewed in Construction Grammar as “stored pairings of form and function” (Gold- 124
berg 2003: 219). It should be clear that there is something incompatible between 125
these two notions, thus defined: on the one hand, pragmatic ways of figuring things 126
out are general heuristics of common-sense reasoning, which means that they are 127
cross-linguistically valid and hence language- and grammar-independent; on the 128
other hand, the constructions of a particular language are conventionalized form- 129
function pairings which make up the language-specific inventory of lexical and 130
grammatical bits of knowledge that speakers and hearers of that language have 131
internalized. So, to the extent that we can more or less agree on these definitions, the 132
question is simply: Can there be such a thing as ‘the pragmatics of a construction’ 133
without this being a contradiction in terms? 134
The second question starts from the assumption that pragmatic information can 135
be found in a construction, for instance in the form of constraints on how it can be 136
used in particular discourse situations. But if this kind of pragmatic information does 137
belong to what is stored in a construction, is it then to be distinguished from other 138
information, in particular ‘semantic’ information? Put differently, if constructions 139
(as we saw) are defined as conventionalized associations of form and function, does 140
that mean that a construction’s semantic and pragmatic specifications jointly make 141
up this functional pole, with no differentiation between them? 142
As for the first question (‘Can pragmatics be part of constructions?’), I will 143
argue that, indeed, there are constructions whose properties must be assumed 144
to include pragmatic information. Some constructions may have interpretations 145
that are purely context-dependent but others may be associated directly with a 146
particular interpretation. We will examine a test by which we can decide whether 147
an interpretation is stored as part of the construction’s properties or not. To lift a tip 148
of the veil of what is to be presented more fully in Sect. 8.3, we will see that we 149
can make a more educated guess about the stored status of a combination’s prag- 150
matically derivable reading by checking whether another language would express 151
that reading in a very similar way. As for the second question (‘Is constructional 152
pragmatics just a kind of constructional semantics?’), we will see in Sect. 8.4 153
that an answer might lie in a model in which stored pragmatic information co- 154
exists with semantic information without coinciding with it. By way of disclaimer, 155
I should state here that this model is not the Construction Grammar solution to 156
integrating pragmatics in constructions. As a matter of fact, there have been many 157
different ways in which construction grammarians have dealt with pragmatics, if 158
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 119
they have dealt with it at all. We will therefore first consider which pragmatic 159
phenomena have been studied in Construction Grammar, and how they have been 160
approached. 161
8.2 Some Constructionist Work on Pragmatic Aspects 162
In a book review on the Linguist List, Ebensgaard (2014) discusses Del Campo 163
Martínez’ (2013) constructionist study of how illocutionary functions (giving an 164
order, making a request, etc.) are encoded in English, and in doing so, he remarks 165
that “systematic studies of pragmatic aspects of constructions are still few and far 166
between”. To some extent, it is true that Construction Grammar, while it proclaims 167
to account for “the rich semantic, pragmatic, and complex formal constraints” on 168
grammatical patterns (Goldberg 2003: 220), has so far not devoted an enormous 169
amount of attention to pragmatics. If we are allowed to give a rather negative 170
appraisal of the state of affairs, it looks as though only few people operating within 171
Construction Grammar consider pragmatics seriously in their actual work, even 172
though not a single construction grammarian would deny that pragmatics plays an 173
important role in many (if perhaps not all) constructions. One indication of this 174
relative lack of interest may be found in the number of page references to the 175
keyword “pragmatics” in the recent handbook of Construction Grammar (Hoffmann 176
and Trousdale 2013): there are only twelve of them, while “semantic” has a hundred 177
page references, almost half of which span multiple pages. 178
Looking at the situation more positively, though, we have to acknowledge the 179
fact that there does exist some work in Construction Grammar that has focused on 180
various ‘pragmatic’ phenomena. Without attempting to define ‘pragmatic(s)’ here, 181
the following overview is meant to give an idea of how construction grammarians 182
have dealt with a range of topics that are widely agreed to come under the purview 183
of pragmatics, as is witnessed by their coverage in general textbooks on this 184
domain of linguistic study: Gricean maxims, information structure, speech acts 185
and discourse patterns. For other overviews, see Kay (2004), Nikiforidou (2009), 186
the introduction to the papers in Bergs and Diewald (2009), and Lee-Goldman 187
(2011). 188
8.2.1 Gricean Maxims in Construction Grammar 189
According to Grice (1975), speakers and hearers expect each other to behave 190
according to certain maxims of conversation. These maxims are stated in very 191
general terms (e.g. “Do not say what you believe to be false”, “Be relevant”, 192
etc.). On the face of it, Gricean conversational maxims could therefore easily 193
be considered to be an ‘add-on’ to grammar, something which is not part of 194
UNCORRECTED
PROOF
120 B. Cappelle
constructional knowledge. Moreover, Grice (1975) himself pointed out that the 195
maxims are equally operative in non-linguistic as in linguistic situations. For 196
instance, obeying the Maxim of Relation means that if you are helping someone 197
in carrying out a task, your physical actions should be appropriate to the needs at 198
hand: “if I am mixing ingredients for a cake, I do not expect to be handed : : : an 199
oven cloth (though this might be an appropriate contribution at a later stage)” (Grice 200
1975: 47). 201
While the maxims are not specifically linguistic in nature, there are conventional 202
expressions in language which are used to negotiate their competing demands, such 203
as giving enough information (Maxim of Quantity) and not saying anything you 204
can’t back up with adequate evidence (Maxim of Quality). One example is the 205
hedge Don’t take my word for it, but : : : . As far as I am aware, only two papers 206
taking a constructionist approach see Gricean pragmatics not as something which 207
interacts with the conventional meaning of a construction (see e.g. Salmon 2015) 208
but describe how a construction can actually contain information for speakers as 209
to which Gricean maxim is being attended to in which part of the construction. 210
The scarcity of this sort of papers is somewhat surprising, in view of the fact that 211
the article which is often considered to be a foundational one for Construction 212
Grammar – the article in Language on the let alone construction by Fillmore et al. 213
(1988) – deals rather explicitly with the distribution of Grice’s (1975) conversational 214
Maxims in this pattern. Consider the exchange in (2): 215216
(2) A: Can the baby crawl yet?B: It can’t even sit up yet, let alone crawl!
217
Given the context proposition in A’s question, the let alone sentence with which 218
B replies reconciles the Maxim of Quantity (by making her conversational contri- 219
bution as informative as is demanded by the exchange) and the Maxim of Relation 220
(by being relevant to the context proposition, which is specifically about the baby’s 221
ability to crawl). The maxims are obeyed in that order: first that of Quantity, then that 222
of Relation. That the first part is more informative is due a sequence of motoric skills 223
that babies generally go through and which B assumes to be common knowledge to 224
A. Since sitting up is acquired more easily and earlier than crawling, it naturally 225
follows that if this more basic skill hasn’t been acquired yet, then any skill that is 226
harder to acquire or comes later in the normal development of babies will not have 227
been acquired either. 228
There is another, less common use of let alone, which has long been overlooked 229
or unjustly rejected as erratic (Pullum 2013) but which Cappelle et al. (2015) argue 230
to be fully legitimate. An example of this alternative use is given in (3): 231
232
233
(3) “Where’s the father of your ten children?”Chelsea looked at me from over her book, “First, we don’t have ten children, let aloneone, and second, he’s outside smoking his life away.”1
234
1http://www.mibba.com/Stories/Read/147313/In-Your-Eyes-I-Lost-My-Place/1/, accessed 3October 2014.
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 121
Compared to the canonical let alone construction used in (2), we can see that in (3), 235
the order in which the maxims of Quantity and Relation are satisfied is reversed. 236
It is the first part of this let alone sentence (we don’t have ten children) that, by 237
denying the context proposition, satisfies the Maxim of Relation. And it is the clause 238
fragment in the second part (whose fully restored proposition is ‘we don’t have one 239
child’) that satisfies the Maxim of Quantity: in our day and age, nobody is surprised 240
to hear that someone does not have ten children, while there is comparatively much 241
more informative value to the statement that someone doesn’t have any children at 242
all. Pullum (2013) disapproves of an authentic example similar to (3) on the grounds 243
that the speaker seems to present the first proposition as semantically entailed by 244
the second proposition rather than vice versa, thereby getting the meaning of the 245
let alone construction all wrong. However, in this alternative use of the let alone 246
construction, the speaker is not at all concerned with making any logical argument. 247
For example, the relevant portion of (3) is different from (4), where the speaker 248
does make the point that if one doesn’t even have one child, it naturally follows one 249
doesn’t have ten children. 250251
(4) We don’t have ten children, fsince/as/given thatg we don’t have (even) one child. 252
So, what motivates the speaker to use the seemingly deviant use of the let alone 253
construction in (3)? Its function appears to be that of providing an afterthought in 254
which the speaker cancels a conversational implicature that could arise from the first 255
proposition. ‘We don’t have ten children’ allows the hearer to infer that the speaker 256
might still have fewer than ten children (i.e. 9, 8, or 7 etc. children). The speaker, 257
having uttered – or while uttering – the first part of the sentence, realizes this and 258
then hastens to add a clause fragment whose (restored) proposition pre-empts the 259
inference of an unwanted implicature in the mind of the hearer. The status of this 260
second utterance as an afterthought is clear from the fact that it is usually possible 261
to insert for that matter after let alone in such cases, as in Boyfriend? She’d never 262
had one, let alone, for that matter, wanted one.2 263
Both Fillmore et al. (1988) and Cappelle et al. (2015) treat the pragmatic 264
information of the construction as distinct from its semantic information. The 265
semantic information states that the second part of the construction, like the 266
first part, must be understood as conveying a full proposition and that these two 267
propositions are in a scalar relationship. The main semantic difference between 268
the two subtypes of the let alone construction is that the first part is the stronger 269
proposition in the canonical type but the weaker proposition in the afterthought 270
type, in terms of noteworthiness. The pragmatics of both the canonical let alone 271
construction and its illogical-seeming afterthought variant specifies the relative 272
degree of relevance and informativeness of the two parts, besides containing some 273
more detailed information that go beyond the mere satisfaction of Gricean maxims 274
in a certain order. This is illustrated here for the type used in (3) (cf. Cappelle et al. 275
2015): 276
2https://www.fanfiction.net/s/2556576/3/Just-a-Girl, accessed 10 October 2014.
UNCORRECTED
PROOF
122 B. Cappelle
‘illogical’ Afterthought let alone Construction 277
Syntax: Xclause let alone Yclause fragment 278
Semantics: Xweaker proposition, and Ystronger proposition 279
Pragmatics: X is more relevant, Y is more informative 280
X could suffice on its own in the context, so it’s not necessary to 281
state Y 282
The speaker states Y to cancel a conversational implicature triggered 283
by X 284
The speaker portrays Y as not worthy of much attention 285
Note that this constructional representation has a tripartite structure, which doesn’t 286
correspond nicely with the popular conception of constructions as pairings of 287
form and meaning (or form and function). However, I find it useful to make a 288
distinction between lexical or propositional semantics, which in the canonical let 289
alone construction further specifies that the second proposition is semantically 290
entailed by the first, and pragmatic information, which encompasses those aspects 291
of a speaker’s knowledge of a linguistic expression that are treated as falling outside 292
the domain of lexical or propositional semantics. 293
Gricean pragmatics in the two let alone constructions discussed here amounts 294
largely to tailoring the sentence according to the demands of the Maxims of Relation 295
and Quantity, by simply obeying these maxims (rather than, for instance, flouting 296
them for special interpretational effects). In fact, this pragmatic aspect of the 297
constructions in essence instructs speakers to distribute old and new information 298
over sentence elements and in that respect is closely related to information structure, 299
a topic we will now turn to. 300
8.2.2 Information Structure in Construction Grammar 301
Quite some constructionist work is concerned with the information status of 302
sentence parts – whether they are topical, focal, presupposed, asserted, weakly or 303
more strongly activated, etc. Indeed, few construction grammarians would disagree 304
with Leino’s (2013: 329) assessment that “[i]nformation structure is without any 305
doubt the pragmatically oriented phenomenon or subject area that has received 306
the most attention in the context of Construction Grammar.” As an example of 307
an information structure construction, consider the Nominal Extraposition (NE) 308
construction discussed by Michaelis and Lambrecht (1996), illustrated in the second 309
sentence of (5): 310311
(5) Announcer: Hear what denture wearers all over America are saying.Denture wearer: It’s AMAZING the DIFFERENCE! [Fixodent commercial](Michaelis and Lambrecht 1996: 216)
312
This construction differs in a number of ways from the more mundane right- 313
dislocation pattern, in which only one element is accentuated: It’s AMAZING , the 314
difference. This latter utterance, but not the one in (5), could be used if the announcer 315
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 123
had said, “Hear what denture wearers all over America are saying about the 316
difference Fixodent has made in their lives”. Crucially, the NE construction has two 317
focal elements: not just the predicate but also the post-predicate NP is in focus. 318
This post-predicate NP refers to a discourse element which, while contextually 319
accessible, is not currently active. 320
Let us now take a look at the Post-Focal comme-N (PFCN) construction in 321
French, meanwhile relabelled as the Right-Detached comme-N (RDCN) construc- 322
tion (Lambrecht 1995, 2004), made use of in the child’s response in (6): 323324
(6) Baby-sitter: Je vais vous raconter une belle histoire, marrante.‘I’m going to tell you a beautiful story, a funny one.’(Starts telling story; child interrupts him.)
Child: C’est pas MARRANT, comme histoire.‘This isn’t a funny story.’
(Lambrecht 1995: 186)
325
The child’s utterance literally means ‘It isn’t funny as story’. The difference between 326
this literal and the freer translation in (6) makes it clear that there is something non- 327
compositional about this pattern: it cannot be strictly predicted from the meaning 328
of its parts that the comme-N phrase should be understood as forming part of 329
the predicate (‘isn’t a funny story’). This interpretation does not follow from the 330
meaning of comme-N outside of the specific grammatical environment c’est AdjP 331
comme N. That the pattern exists at all and has the form that it does is argued by 332
Lambrecht to be motivated by the need to disambiguate another pattern in French: 333
C’est pas une historie MARRANTE , which could either mean something like ‘What 334
one can say about this story is that it isn’t a funny one’ or ‘What one can say 335
about this is that it isn’t a funny story’. The Post-Focal comme-N construction is not 336
similarly ambiguous. In (6), it is perfectly clear to both the child and the babysitter 337
that a story is being told. In this respect, this construction is also very different 338
from the Nominal Extraposition in (6), where the extraposed NP does not denote 339
something presupposed in the discourse. 340
As with Leino’s (2013) earlier quote above, it would again be hard to find 341
anything unfair in his statement that “[i]n the context of and in relation to Construc- 342
tion Grammar, Information Structure has been most studied by Knud Lambrecht 343
[references omitted]” (Leino’s (2013: 329). An indispensable book-length construc- 344
tionist treatment of information structure is Lambrecht’s (1994) monograph, which 345
discusses crucial notions such as topic, focus, pragmatic presupposition, etc. as 346
they play out in particular sentence types. For Lambrecht, information structure 347
constitutes a third layer in the architecture of grammar, alongside (morpho)syntax 348
and semantics, so his work, like Fillmore et al. (1988), diverges from the now- 349
prevailing dichotomous conception of constructions. Lambrecht may be the most 350
important scholar to have looked into information-structural properties of construc- 351
tions, but he is not the only construction grammarian that can be mentioned in 352
connection with information structure. For instance, Goldberg (2000, 2004) shows 353
how the Deprofiled Object Construction allows patient arguments normally required 354
by the verb to be omissible when they have low discourse prominence and when the 355
action is emphasized (e.g. The singer always aimed to dazzle), an analysis to which 356
UNCORRECTED
PROOF
124 B. Cappelle
Lemmens (2006) provides some refinements. Ambridge and Goldberg (2008) argue 357
that Ross’s (1967) ‘islands’ for extraction involve backgrounded information. Thus, 358
in (7), a light verb such as say or think presents the contents of its complement as 359
asserted, unlike a manner-of-speaking verb such as mumble or a factual verb such 360
as realize, whose complement provides backgrounded (presupposed) information: 361362
(7) Who did she fsay/think/??mumble/??realizeg that he saw __ ? 363
Kanetani (2009, 2012) discusses Metalinguistic Because Constructions, first noted 364
by Hirose (1992), which take an expression from the preceding context and justify 365
their use by the (current or original) speaker. The attested examples in (8) provide 366
an illustration: 367
368
369
(8) a. Cyber bullying in the Philippines is a growing threat. “Growing”, because internetpenetration rate is increasing and access is becoming easy. (GloWbE)
b. The battles over all this will, to some extent, dominate our politics henceforward.We got a glimpse of the nature of the fight over the debt ceiling in July, and the2012 election will pivot on it. I say “to some extent” because unexpected events,probably in the realm of foreign policy, will surely come along to complicate thepicture. (COCA)
370
These constructions come in two semantically equivalent types, one with and 371
another without an initial performative like I say (or sometimes They say) as an 372
‘activator’, the choice between these types arguably depending on how accessible 373
the speaker assumes the word or expression in question still is to the hearer: 374
note how in (8b), to some extent was used not in the previous sentence but in 375
the sentence before it. Deulofeu and Debaisieux (2009) deal with a somewhat 376
similar construction in spoken French, the Context Focussing Parce Que (‘because’) 377
Construction, which is syntactically not an adjunct and can be inserted at any place 378
in a host utterance. An example is provided in (9): 379
380
381
(9) le roi parce que le roi est très aimé en Thaïlande je crois que j’en ai parlé avant le roiest arrivé [ : : : ] : : : à leur faire comprendre que cette culture était très en dent de scie‘the king’cos the king is very much loved in Thailand – I think I have said thatbefore – the king succeeded in making them understand that this culture was veryuneven’ (Deulofeu and Debaisieux 2009: 44)
382
The pragmatic function of this construction is either to bring the requisite back- 383
ground knowledge into the common ground or to reframe the ground so that the 384
hearer can maximize uptake of information. 385
As a final example, consider the transitive Verb-Particle Construction, where 386
the choice between the ‘joined’ order (e.g. pick up the kids from school) and the 387
‘split’ order (e.g. pick the kids up from school) is influenced not only by lexical 388
and semantic preferences but by a variety of context-related ones, such as the 389
(discourse-)familiarity of the object NP, as discussed by e.g. Gries (2003) and 390
Cappelle (2009). Familiarity goes beyond what has been introduced in the discourse 391
or what the hearer is expected to know already; it also includes how ‘normal’ or 392
‘routine’ a situation is (cf. Bolinger 1971), which helps us understand why for 393
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 125
example websites on “how to pick up a girl” vastly outnumber those on “how to pick 394
a girl up”, something which the discourse-newness of the object NP alone might not 395
be able to explain. 396
For more general introductions to information structure in Construction Gram- 397
mar, I refer the reader to Leino (2013) and Hilpert (2014: Ch. 5). What I hope 398
the examples above have illustrated sufficiently is that there are constructions 399
that put constraints on the information status of its components. This makes 400
these constructions context-sensitive. Yet, the context-sensitivity of a construction 401
does not mean that its interpretation is necessarily variable and unpredictable. 402
On the contrary, since the constraints are constant, such a construction leads to 403
similar interpretations, whenever it is used, about what part of the construction is 404
represented as new, focal, backgrounded, etc. 405
8.2.3 Speech Acts in Construction Grammar 406
Lakoff (1987: 473ff) distinguishes a number of what he calls speech act con- 407
structions, including deictic there constructions, negative questions, inverted excla- 408
mations and wh-exclamations, and discusses their (in)compatibility with certain 409
subordinate clauses. For instance, the inverted exclamation Am I ever hungry! 410
indirectly conveys the statement ‘I’m hungry’ and is therefore permitted in so-called 411
performative subordinate clauses. Compare: 412413
(10) a. I’m gonna have breakfast now, because am I ever hungry!b. *I’m gonna have breakfast now, if am I ever hungry!
414
This contrastive pair is clear evidence that sentence-final because-clauses are 415
performative subordinate clauses – they are used to perform a speech act that by 416
convention conveys a statement and at the same time to provide this as a reason for 417
the main clause statement – while if -clauses are not. 418
An often-cited example of a special construction with a stable illocutionary force 419
is what Akmaijan (1984) refers to as the ‘Mad Magazine’ sentence type, after 420
the famous sentence What – Me worry? of Mad Magazine’s cover boy Alfred E. 421
Neumann. An illustration is given in (11): 422423
(11) Him write a thesis in neuroscience?! (Yeah, sure!) 424
This sentence type is also known as the ‘Incredulity Response Construction’. As 425
Lambrecht (1990) argues, it quotes the Subject and (part of) the Predicate of a 426
context sentence to express the speaker’s disbelief in its propositional content, or 427
at least to express the speaker’s evaluation of the juxtaposition of the Subject and 428
the Predicate as something quite incongruous (Szczesniak and Pachoł 2015). 429
Del Campo Martínez and Ruiz de Mendoza (2012) deal with diverse construc- 430
tional realizations of orders. For instance, using the interrogative construction Can 431
you XVP?, which is essentially a request construction, is one of several ways of 432
UNCORRECTED
PROOF
126 B. Cappelle
issuing an order. Thus, an utterance like Can you shut up for a minute? will be 433
understood as an order by virtue of a social convention that stipulates that “[i]f it 434
is manifest to A that a particular state of affairs is not beneficial to B, and if A 435
has the capacity to change that state of affairs, then A should do so” (Del Campo 436
Martínez and Ruiz de Mendoza 2012: 16). A range of similar cultural principles 437
about social interaction make up the cost-benefit cognitive model, which capture 438
speakers’ very general knowledge that they are supposed to help each other if it is 439
at all within their power to do so. Del Campo Martínez’ (2013) study, mentioned at 440
the beginning of Sect. 8.2, relies on this high-level cognitive model of social norms 441
to account for conventionalized ways of realizing a wide range of speech acts: those 442
of ordering, requesting, advising, offering, promising, threatening, congratulating, 443
thanking, apologizing, pardoning, condoling and boasting. 444
Stefanowitsch (2003) analyses the Can-you-X Construction and the Would- 445
you-mind-X Construction. He introduces an Illocutionary force (If) parameter 446
(or ‘attribute’ in Construction Grammar parlance) in these indirect speech act 447
constructions. This parameter is filled with the conventionalized value of ‘request’ 448
but is left empty in the ‘direct’ constructions they inherit from, in which can means 449
‘be able to’ and would mind means nothing more than ‘object to’. We will come 450
back to this proposal in Sects. 8.3 and 8.4. 451
Cappelle (2003) discusses a construction with the particle dè in West-Flemish, 452
which, rather like the What’s X doing Y? construction we considered in the 453
introduction, can signal the speaker’s surprise at the situation expressed and requests 454
the hearer to provide some explanation for it (Haegeman 1993). Let’s consider the 455
utterance in (12): 456
457
458
(12) Ah, Maurice, zie j’hier ook dè? (On bumping into the hearer at a different placethan usual)ah, Maurice, are you here too Prt‘Why, Maurice, you’re here too!’
459
The West-Flemish sentence in (12) is interrogative in form but it is obvious that 460
its function is not to ask whether the hearer is really present – that much can be 461
seen directly. The expression of surprise is not conveyed by the dialogic discourse 462
particle dè as such. Rather, this particle is associated with different illocutionary 463
functions in different sentence types (closed interrogatives, wh-interrogatives with 464
rising vs. falling intonation, declaratives, imperatives, elliptic sentences). Thus, in 465
an utterance such as Oe noemt ie were dè? ‘What’s his name again Prt?’, there is no 466
surprise expressed at all. Instead, this wh-interrogative with falling intonation and 467
dè conventionally signals that the speaker has a lacuna in his memory and prompts 468
the hearer to help him search for the forgotten word, name, song, etc. The diverse 469
sentence types with dè have to be stored as distinct, if related, dè-constructions, 470
since their discursive properties cannot be predicted from any known properties of 471
the sentence types combined with any unitary functional description of dè. 472
Fried and Östman (2005) provide a description in much the same vein of some 473
dialogic particles in Solv and Czech. They argue that these particles require detailed 474
and systematic description in interactional terms, which Construction Grammar is 475
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 127
able to capture if it is enriched with a set of parameters – or ‘attributes’ – which 476
are to be filled with a wide range of pragmatic information – or ‘values’ (cf. also 477
Wide (2009) for a similar proposal). These pragmatic parameters include the type 478
of speech act (question, request, assertion, etc.), speaker information (male/female, 479
younger/older, etc.), specification of whether or not the particle, as used in a 480
particular sentence type, marks a shift of discourse topic, whether or not it expects 481
the hearer to give a more or less specific response, how formal or informal it is felt 482
to be, whether it conveys distance, deference or camaraderie, whether the speaker 483
is positively or negatively involved, and so on. For Fried and Östman (2005), the 484
notion of construction should thus be extended in a dialogic direction. Construction 485
Grammar (CxG) is then in line with one of the major approaches to spoken language, 486
namely Conversation Analysis (CA): “The fundamental feature shared between 487
CxG and CA is the fact that both approaches take all ‘chunks’ of language – large 488
and small, verbal and non-verbal – as equally deserving of description and analysis, 489
without assuming any a priori determined relative degree of theoretical significance 490
that a particular type of structure might have for our understanding of grammatical 491
knowledge” (Fried and Östman 2005: 1754–1755). 492
In Construction morphology (Booij 2010), some specific morphological con- 493
structions have been shown to have unexpected speech act meanings. Coussé and 494
Oosterhof (2012) discuss past participles in Dutch which have the same value 495
as an imperative (e.g. opgerot! ‘bugger off!’, ingerukt! ‘march off!’, opgepast 496
‘be careful!’, niet getreurd! ‘don’t be sad!’). Fortuin (2003) and Van Olmen 497
(2009, 2010), inter alia, discuss cases in which the Dutch infinitive, too, has 498
directive force (e.g. Even kijken ‘Let’s see’, Niet op het veld lopen! ‘Don’t run 499
on the pitch!’). Evaluative affixes can have a wide variety of often contradictory 500
expressive meanings (Jurafsky 1996). For instance, Fortin (2011) points out that 501
the Spanish (so-called) diminutive affix -it- may express deprecation (e.g. mujer- 502
cita ‘woman’Ccontempt) as well as appreciation (e.g. niñ-ito ‘boy’Caffection), 503
intensification as well as attenuation (e.g. ahor-ita now-DIM ‘immediately, right 504
now’ or ‘soon, in a little while’), and exactness (e.g. igual-ito ‘exactly the same’) as 505
well as approximation (e.g. mal-ito ‘kind of bad’) and that, when used in requests, 506
it has the pragmatic effect of illocutionary mitigation, lowering the degree of 507
imposition on the addressee (e.g. ¿Me harías un sangüich-ito, porfa? ‘Could you 508
rustle me up a quick sandwich, please?’). In Construction Grammar, evaluative 509
meanings of affixes and affixoids have recently been discussed by Booij and 510
Hüning (2014), Morris (2013), Norde and Morris (2014), Trousdale (2014) and Van 511
Goethem and Hiligsmann (2014). 512
From the above, it appears that there are quite a few constructions which do 513
not just indicate that the utterance is to be understood as a statement, a question, 514
a command, etc. but which also, or especially, betray the speaker’s heightened 515
emotional involvement in making the utterance. This expressive dimension (cf. 516
Potts 2007) has turned out to be vital for the description of a range of further 517
constructions, such as phrasal verbs (Hampe 2002, Cappelle 2005) and various 518
cases of so-called insubordination (illustrated by As if I would ever do that! or 519
Well, darling, that you of all people should say that! (L.P. Hartley)), which have 520
UNCORRECTED
PROOF
128 B. Cappelle
recently been studied for a variety of languages (cf., inter alia, Boogaart 2015, 521
Brinton 2011, D’Hertefelt and Verstraete 2014, Kaltenböck 2006, Schwenter 1996, 522
2016, Van linden and Van de Velde 2014). 523
8.2.4 Discourse Patterns in Construction Grammar 524
Given what we have quoted in Sect. 8.2.3 about the link between Construction 525
Grammar and Conversation Analysis, it will not come as a surprise that some 526
linguists – most notably Jan-Ola Östman – have put forward the idea that larger 527
discourse patterns, exceeding the length of a single utterance, can be described as 528
conventional constructions (Östman 2005). Some of the very specific text types that 529
have thus been approached from a constructionist point of view are news reports 530
(Östman 1999), postcards (Östman 2000), and football chants (Hoffmann and Bergs 531
2012). Bergs (2008) suggests that text types and genres can be integrated in a 532
structured inventory of interlinked constructions, known in Construction Grammar 533
as the ‘constructicon’: 534
Text types such as letters, postcards, faxes, telegrams, emails, and short messages are 535
related and connected in some specific ways, and so are novels, short stories, and other 536
literary genres and subgenres. Also, some genre-constructions actually comprise a number 537
of subconstructions. Newspapers, complex and fairly schematic, are organized and layouted 538
in very specific ways; they include the editorial, the sports sections, finance etc., i.e. less 539
complex but still quite schematic subparts. In Aristotelian drama there are three different 540
main parts or, according to Freytag, five different parts; the Shakespearean sonnet comprises 541
fourteen lines and balances quatrains and tercets (or quatrains and the final couplet); stories 542
of adventure have recurring episodes constructed according to basic narrative structures. 543
All these elements need to come together in order to result in the complex construction of 544
a certain text. A novel without a title would be awkward, and so would a recipe without 545
ingredients; a sonnet with fifteen lines would strike readers as remarkable, and so should a 546
newspaper without headlines. (Bergs 2008: 274) 547
While proposals to accommodate such genre-level or text-type constructions in a 548
Construction Grammar framework are still sketchy, the potential is there and the 549
idea is definitely worth further exploring. 550
8.2.5 Interim and Transition 551
The impression one may get from the above brief survey is that in Construction 552
Grammar research, we find a rich diversity of pragmatic phenomena that have been 553
incorporated in the description and representation of constructions. Among these 554
pragmatic aspects of constructions, we have discussed obedience of constructions’ 555
parts to Gricean maxims, (other) information-structural properties, knowledge 556
about constructions’ illocutionary force and about their function in interaction and 557
knowledge of how constructions fit into larger discourse patterns – or, conversely, 558
of how such larger discourse patterns, viewed as constructions, can be broken 559
8.2.5 Interim and Transition
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 129
down into smaller constructions. We have not even discussed broad cross-cultural 560
differences, but these too could in principle be given their place within an enlarged 561
constructicon. As Yukio Hirose (p.c.) sees it: 562
I understand some linguists like Hans Boas speak of “Contrastive Construction Grammar” 563
[cf. Boas 2010 – B.C.]. If Contrastive Construction Grammar is a meaningful endeavor, as I 564
believe it is, then it should incorporate the general differences (whether typological or not) 565
between languages, because particular languages such as Japanese and English have their 566
own default preferences in construal and the mode of expression, which are expected to be 567
inherited directly or indirectly by particular constructions in those languages. This way of 568
thinking, I believe, is in keeping with the spirit of Construction Grammar. 569
While there is no shortage of studies of specific pragmatic phenomena – although, 570
of course, the more the better – my feeling is that there is an urgent need for some 571
more theoretical reflection about what kind of pragmatic information should and 572
should not be included in constructions and how, if at all, pragmatics differs from 573
semantics. I will address these issues in the remainder of this chapter. 574
8.3 Storage vs. Online Computation 575
Let’s consider the following fundamental question formulated by Jackendoff 576
(2002): 577
What aspects of an utterance must be stored in long-term memory, and what aspects can be 578
constructed online in working memory? (Jackendoff 2002: 152, emphases in original) 579
I will not go into the reasons why Jackendoff asks this question, repeated in 580
Jackendoff (2010: 226), or into the way he elaborates on it. I merely cite it here 581
because it seems to me to lie at the heart of Croft’s (1998) continuum of possible 582
mental representations of lexical or grammatical knowledge, simplified in Fig. 8.1. 583
At the extreme left of the cline, we find the independent entries model, where 584
two formally identical items (or one item and another formally related one) are 585
stored – or entrenched – in the mind as distinct form/meaning pairings. The 586
mind has routinized the link between these forms and their respective meanings. 587
separate storage of
individual form-
use pairings:
[a/U1] and [a/U2]
(or [ab/U2])
storage of just the most
general meaning of forms
but not of their derivable
uses: [a/U] (and [b/W])
but not (a/U1) and (a/U2)
(or (ab/U2))
Fig. 8.1 A cline of possible mental representations of form a with two senses U1 and U2 (or ofform a on its own with sense U1 and of form combination ab with sense U2, where b on its ownhas meaning W); mentally stored items are represented between square brackets
UNCORRECTED
PROOF
130 B. Cappelle
Consequently, language users do not need to work out what uses these forms could 588
possibly have (although in the case of, for example, bank1 ‘financial institution’ and 589
bank2 ‘slope of land along a river’, the hearer will still have to decide which of the 590
stored meanings has to be retrieved from the lexicon in a particular usage situation 591
such as Let’s meet at the bank). At the opposite end of the cline, to the extreme 592
right, the model of representation can be called the pragmatic model, but I prefer the 593
alternative term that Croft (1998) suggests, namely discourse model, as it stresses 594
the importance of context-sensitive interpretations of words as used in particular 595
discourse environments. Here, only one general meaning of a word or expression 596
is stored. The more specific uses of this form (on its own or in combination with 597
another form) are not similarly entrenched; instead, language users compute these 598
uses with the aid of general (language-independent) cognitive principles applied to 599
ever-changing contexts of use. 600
We might think that the availability of such ‘pragmatic’ rules, in the sense of 601
general principles of common-sense reasoning and problem-solving, spells nothing 602
good for the constructional status of an expression, that is, for its status as a stored 603
unit. After all, why would we have to store an expression in the lexicon if its correct 604
interpretation may be obvious even to a hearer who didn’t know the expression (and 605
if, likewise, coming up with this expression to convey that interpretation could be 606
an obvious choice for a speaker who didn’t already have that expression stored in 607
the mind)? Yet, storage may still be a possible scenario in that case. Croft points 608
out that while we can use purely linguistic – by which is meant introspective – 609
knowledge to exclude one or more general models on the continuum (i.e., models 610
towards the right in Fig. 8.1), we can’t rely on such knowledge to exclude the more 611
specific models (i.e., the ones towards the left). On the one hand, it’s clear that 612
form-use units must be stored if more general representation models are excluded 613
on the basis of introspective evidence for these units’ irregularity and idiosyncrasy: 614
if there’s no rule by which they can be interpreted, their meaning will have to be 615
learned as such (or, from the perspective of the speaker, choosing a form to convey 616
that interpretation will also have to be part of explicit language knowledge). But 617
on the other hand, it should also be clear that just because there is introspective 618
evidence for regularity or generality of a form-use pairing, this doesn’t mean that 619
it is always (or ever) constructed by applying this rule. So, form-use units which 620
fit a rule can of course be generated by this rule, but they don’t have to and in fact 621
may not be. This explains why Jackendoff (2002: 152) uses the qualifications must 622
and can in the question above: Which parts must be stored and which parts can 623
in principle be constructed on the fly even though they may not be? As Langacker 624
(1987) convincingly argues, it would be erroneous to assume that an expression that 625
is grammatically well-behaved cannot, ipso facto, be stored as a fully memorized 626
entity. This mistaken assumption – that expressions either follow rules or have to be 627
listed as exceptions to the rule-system – is known as the ‘rule-list fallacy’. One of 628
the things that psycholinguistics has taught us is that entities that are very frequent 629
in use are likely to be stored as such, regardless of whether or not they are fully 630
regular. 631
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 131
To summarize, by using purely linguistic methods of analysis, we can prove that 632
an item is ill-behaved (i.e., non-compositional) and so must be locked up, so to 633
speak, in the constructicon; but we cannot prove with purely linguistic methods 634
that a well-behaved (i.e., compositional) item is not stored as an entry in the 635
constructicon. 636
8.3.1 Excluding the Discourse Model of Mental 637
Representation: An Example 638
So, what can we do as ‘ordinary’ linguists, without having to become psy- 639
cholinguists or neurolinguists? As Croft (1998) demonstrates, for any form-use 640
association, we can exclude the extreme discourse model as soon as we find that this 641
association is less than fully predictable. Remember that in this model the form-use 642
pair is pragmatically derived and is not stored in the mind. Obviously, if the form is 643
somehow irregular or if the interpretation is not fully compositional, the item needs 644
to be learned as part of the conventions of the language. Croft (1998) proposes 645
a fairly straightforward means of excluding the discourse model for any given 646
form-use pair. The linguist needs to check whether the translation equivalent in 647
another language is formally equivalent (taking into account non-pertinent structural 648
differences as well as trivial lexical differences). If there is no formal equivalence, 649
then the use of that supposedly pragmatically derived unit must be considered to be 650
stored as such in the first language, rather than being computed on-line. 651
Looking at translation equivalents for the purpose of excluding (or substantiat- 652
ing) the discourse model is a procedure which Lee-Goldman (2011) also considers 653
an indispensable one: 654
Cross-linguistic examination is crucial for [the] study of the grammar-context interface. 655
It is possible to argue purely within English that This is Kim [as used to identify oneself 656
on the phone – B.C.] instantiates a special construction, but there is always the question: 657
could its use not simply be a matter of non-linguistic reasoning and inference, such as 658
from Gricean maxims? Cross-linguistic comparison can defeat (or support) this objection: if 659
another language has similar linguistic parts at its disposal but nonetheless cannot combine 660
them in the same way, with the same interpretation or function, then it is doubtful that 661
inference alone will account for the data. (Lee-Goldman 2011: 8–9) 662
Let’s apply this to water tower, a case discussed by Croft (1998: 158–159). This 663
word is a noun-noun compound which essentially refers, as its name suggests, to a 664
tower containing water. Surely this form needn’t be stored, for what else could such 665
a thing be called? The thing is, however, that such a structure is referred to in French 666
as a château d’eau ‘water castle’ and not as a tour d’eau ‘water tower’. Note, by the 667
way, that English [nouna nounb] compounds correspond with French structures of 668
the form [nounb de nouna], but this basic difference is ignored in checking formal 669
equivalence. Given the lack of formal equivalence, Croft (1998: 159) concludes: 670
“This cross-linguistic difference is strong evidence that the pragmatic model is not 671
appropriate for the derivation of this particular noun compound; otherwise speakers 672
UNCORRECTED
PROOF
132 B. Cappelle
of both languages would come to the same solution to the naming of this object”. 673
Critics might object that calling this object a water castle, as the French apparently 674
do (but then in French of course), is a rather marked choice in that language and 675
therefore needn’t force us to conclude that the compound water tower in English is 676
not a fully transparent combination that could be constructed online. Or they might 677
object that even for speakers of French, calling that object water castle is the most 678
natural choice given the abundance of castles in their non-linguistic environment and 679
given the physical similarity of the earliest water towers to castle towers. But then 680
again, when the first water towers appeared in the UK and the US, many of these 681
exemplified the then-popular Gothic revival architecture no less than in France. In 682
other words, a water tower could also have been called water castle in English (or 683
perhaps water fortress or water keep). Other imaginable alternatives could have been 684
water house (by analogy of light house), water storage tower, water supply tower, 685
hydrotower, water building, water pillar, water silo, water tank tower, cistern tower, 686
at least some of these terms being literal translations of terms for this structure in 687
other languages (e.g. in Japanese). In fact, an alternative that is actually used is 688
standpipe. In short, alternative encodings in other languages – or in (older stages 689
of) the same language – may help us realize that a complex expression which seems 690
transparently compositional is not the only obvious choice (as a standard encoding 691
option) and thus may have to be listed as a learned language item. We can make a 692
further claim here: once a combination is conventionalized, it is no longer assembled 693
from scratch each time it is used, nor is it decoded inferentially each time the hearer 694
comes across it. What I mean, to put things more simply, is that a native speaker will 695
not ask, “What should I call this tower-like thing used to get pressure on water”?” 696
and then say, “Hmm, I guess I’ll call it a water tower.” And a hearer competent in 697
English will not ask, “Water tower? What does that mean?” before concluding that 698
it definitely can’t be a tower made of water, so that it must be a tower containing 699
water. 700
The example given here may raise many questions. As I have already hinted 701
at, a main concern is how we can maintain that the existence of a formally non- 702
equivalent translation (e.g. château d’eau in French) is proof that the expression 703
under discussion (water tower in English) is non-compositional, while it may be 704
actually the translation that is non-compositional. After all, a château d’eau is 705
certainly not a castle, while a water tower is definitely a tower. To decide this 706
issue, we could perhaps look at several other languages. Thus, we might find that 707
Dutch and German have watertoren and Wasserturm, respectively, which literally 708
also mean water tower. However, does this then mean that water tower in English is 709
compositional after all? Not necessarily, as several languages might have stumbled 710
upon the same not-fully-compositional naming solution for an object – or have 711
borrowed this solution from each other. In any case, the procedure of looking at 712
another language and finding something rather different there can helpfully alert 713
us to the possibility (without this being a certainty) that the expression in the first 714
language is less obvious than meets the eye. 715
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 133
8.3.2 Excessive Resultatives (Talk One’s Head Off, Sing One’s 716
Heart Out, Cry One’s Eyes Out, etc.) 717
We can apply the cross-linguistic method just demonstrated to more productive 718
patterns as well. Let us take the Body Part Off /Out Construction (BPOC) (e.g. 719
Jackendoff 2002: 173–74; Glasbey 2003; Mateu and Espinal 2007; Espinal and 720
Mateu 2010; Kudo 2011). In Kudo’s (2011) original description of this pattern (later 721
on revised in his PhD thesis), a discourse model of representation was proposed for 722
instances of this pattern. The reason seemed straightforward. Suppose we hear the 723
utterance Jane cried her eyes out, then, according to Kudo (2011), we first process 724
the literal interpretation. We then realize that, obviously, nobody literally causes his 725
eyes to pop out of his head as a result of crying, so we finally infer that the utterance 726
must be understood as a hyperbole. Cappelle (2014) rejects this pragmatic model 727
of mental representation, using the diagnostic shown for water tower above. Let’s 728
consider the steps in the reasoning. 729
First, it can be observed Dutch, which has property and path resultatives much 730
like English does, lacks a perfect equivalent of the BPOC. That is to say, the closest 731
expression is one with a full PP, whose word-by-word translation in English would 732
sound like cry one’s eyes out of one’s head; sing one’s lungs out of one’s body, run 733
one’s legs off of one’s body, etc. Yet, what is even more common in Dutch is an 734
excessively used (reflexive) ditransitive, where the Theme argument often denotes a 735
case of physical damage or deformation, which in English would sound like laugh 736
oneself a fracture, sweat oneself an accident, search oneself a hunch, or where the 737
Theme argument is a nasty disease, usually one that is no longer common but still 738
lives on in public memory, translating as work oneself the pleurisy, startle oneself 739
the typhus, bore oneself the plague, or occasionally involves an imaginary disease 740
that only occurs in this kind of intensifying language use, as in zich het apelazerus 741
werken (lit. ‘work oneself the monkey leprosy’). For an overview of Themes used in 742
this pattern, see Morris (2014). 743
Second, given this formal non-equivalence, we cannot just assume that the BPOC 744
in English constitutes a set of expressions that are simply to be expected in the 745
language. Even if their interpretation may pose no problem, the speaker still has to 746
learn that in English one says run one’s legs off and not run one’s legs off of one’s 747
body or run oneself the pleurisy. Again, we see that more is stored in a language 748
than one might at first sight assume. Even though the form of an expression such 749
as cry one’s eyes out and its interpretation could in principle arise from processes 750
involving implicature/inferencing, cross-linguistic evidence casts doubt on whether 751
this is actually the case. In fact, the cross-linguistic test applied here forces us to 752
abandon the computational model of representation. 753
UNCORRECTED
PROOF
134 B. Cappelle
8.3.3 Can You X? 754
Among the most widely used examples of indirect speech acts (ISAs) in the 755
literature, we certainly find Can you X?, as in Can you pass the salt? or Can 756
you close the door? By itself, that is, out of context, this latter sentence could 757
mean two things: (i) ‘Are you able to close the door?’ and (ii) ‘(I request that 758
you) close the door’. Thus, it could be either a question, when it is uttered as a 759
direct speech act, or a request, when uttered as an indirect one. As we saw above in 760
Sect. 8.2.3, Stefanowitsch (2003) treats Can/Could you X?, used to express requests, 761
as a separate construction, a stored unit: 762
[T]he ISA has construction status in spite of the theoretical possibility to derive its 763
meaning from the request scenario on-line (as must in fact be done in the case of non- 764
conventionalized ISAs). (Stefanowitsch 2003: 117) 765
So, why exactly is Can you X? a construction? Can’t the request meaning be worked 766
out via Gricean principles? The answer, according to Stefanowitsch (2003), lies in 767
the observation that some of its formal aspects are unpredictable. We can apply a 768
language-internal diagnostic similar to the cross-linguistic formal non-equivalence 769
test illustrated above for water tower and for the Body-Part Off /Out Construction. 770
In this case, what we can do is compare Can you X? with one or more semantically 771
equivalent expressions in the same language, and check whether these expressions 772
have similar or different formal properties. Can you close the door? (D13a) has 773
as some of its close alternatives (13b) and (13c), at least when it is intended as a 774
request: 775776
777
(13) a. Can you close the door?b. Are you able to close the door? (semantically close to (13a) in meaning (ii)
above)c. Is it possible (for you) to close the door? (idem)
778
Now, as had been noted by Sadock (1974: 90), only (13a) but not (13b-c) allows 779
sentence-internal addition of the illocutionary force indicating device please: 780781
(14) a. Can you please close the door?b. *Are you able to please close the door?c. *Is it possible (for you) to please close the door?
782
So, in its request meaning, (13a) apparently resembles a direct request such as 783
Close the door, please, while (13b) and (13c), though equivalent to (13a) in one 784
of its meanings, resemble non-conventionalized indirect speech acts, which do 785
not license the insertion of please either (e.g. *It’s cold in here, please). It is an 786
unpredictable fact of Can you X? that, unlike its close equivalents, it supports the 787
pre-verbal addition of please. This fact requires storage of this form together with its 788
illocutionary force, which allows it to be directly interpreted as a request, so that we 789
‘know’ it can have please, since this is the case with structures directly associated 790
with the illocutionary force of request. 791
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 135
We have to be careful in how we formulate our line of argumentation. It’s not the 792
case that (13b) and (13c) cannot be used as indirect speech acts for making requests, 793
as Sadock (1974: 78, 90) claims. This is clear from example (15a) given by Brown 794
and Levinson (1987: 139), for which there is ample evidence in actual language use, 795
and a corpus-attested example such as (15b) provided by Leech (2014: 306): 796
797
798
(15) a. Are you by any chance able to post this letter for me? (Brown and Levinson1987: 139)
b. Is it possible to lay the table, do you think, for me? (Leech 2014: 306)799
In other words, what we find is not that Can you X? has an unpredictable semantic 800
property, but rather that it has an unpredictable formal property (viz., pre-verbal 801
insertability of please). Either of these unpredictable properties suffices to grant a 802
sequence construction status. 803
8.3.4 On So-Called ‘Pragmaticalization’ and ‘Short-Circuited 804
Implicatures’ 805
For Leech (2014), the likely reason why cases like (14b) (and, similarly, (14c)) are 806
not found in corpora is this: 807
Are you able : : : ? is not sufficiently “routinized” or pragmaticalized [ : : : ] as a directive 808
formula in English to be able to co-occur with the overt directive pragmatic marker please. 809
(Leech 2014: 305–306; italics in original; cross-reference omitted from the citation – B.C.). 810
The term pragmaticalization was first introduced by Erman and Kotsinas (1993); 811
see also, inter alia, Aijmer (1997), Claridge and Arnovick (2010), Diewald (2011), 812
and Lauwers et al. (2012). 813
Note that pragmaticalization as a term may be confusing. It certainly does not 814
mean ‘development whereby more room is given to language-independent processes 815
of inferential reasoning’. On the contrary, it refers to the conventionalization of 816
what was once a pragmatic interpretation, where I use pragmatic in the sense of 817
‘going beyond the literal, coded meaning’. This is why we can speak of an actual 818
construction in the case of Can you X?. We are dealing here with a form-function 819
pairing stored in the mind. The form no longer has to be constructed by the speaker – 820
nor does the intended meaning have to be calculated by the hearer – online. The 821
form-function association is ready-made. 822
A related concept in this context is that of a “short-circuited implicature”, coined 823
by Morgan (1977: 23), that is, an implicature which the hearer arrives at directly, 824
without having to calculate it. Morgan (1977: 23 ff.) provides many telling examples 825
of fixed and clichéd expressions, all of which have an implicature that is somehow 826
conventionalized. Here is one example discussed by him: 827
There are many cases that involve implicature, but as a matter of convention, so that the 828
implicature is short-circuited. For example, (16) is commonly used to convey something 829
like (17). 830
UNCORRECTED
PROOF
136 B. Cappelle
(16) If you’ve seen one, you’ve seen them all.(17) They’re all alike, so it’s a waste of time to examine them separately.
831
It is intuitively clear that (17) could be reasoned out as an implicature of (16); but it is 832
now conventional to use (16) to convey (17). Although one has in mind the literal meaning 833
of (16) in using it to convey (17), the form of the expression is strictly part of the convention. 834
Sentences having precisely the same literal meaning but even slightly different form do not 835
convey (17) with the same immediacy. If one manages to convey (17) by saying (18) or 836
(19), it will be as a fresh implicature, without the short-circuiting that accompanies the 837
conventional form (16). 838
839
(18) If you’ve seen one, you’ve seen all of them.(19) You’ve seen them all if you’ve seen one.” (Morgan 1977: 26; original numbering of
examples adapted to fit the present text)840
Morgan discusses many similar cases, some of which are given in (20): 841
842
843
(20) a. You can say that again. (cp. You can repeat that.)b. How many times have I told you : : : (cp. Tell me how many times I’ve told
you : : : )c. It takes one to know one. (cp. It requires one to recognize one.)d. [Policeman to motorist] Where’s the fire?e. [Spouse to spouse] I’ve got a headache.f. Your place or mine?g. Are you crazy?/Have you lost your mind?/Are you out of your gourd?h. Is the Pope Catholic? [as an answer to an obvious question] (Morgan 1977: 27 ff.)
844
Language is full of such clichéd sentences. The interesting thing about them is that 845
they are not quite idioms, since they still have a literal meaning which could in 846
principle function as the basis for an inferential process. This inference, however, 847
has come to be by-passed and the erstwhile indirect reading is now a direct one. 848
All of these sentences instantiate constructions with pre-installed pragmatics. In the 849
next section, we will address the question whether we can give both pragmatics and 850
semantics a place in constructions without having to reduce either of them to the 851
other. 852
8.4 Semantics and Pragmatics 853
8.4.1 More on Can You X? 854
In discussing Can you X?, Morgan (1977) writes: 855
the request nature of the speech act is conveyed without the sort of indirect feeling we 856
attribute to the presence of inference; the literal meaning is in some way latent, rather than 857
the basis for an inference. (Morgan 1977:23) 858
Observe that the literal meaning is considered by Morgan to be part of the pattern, 859
although it is present only in the background. This does not seem to be the 860
case in Stefanowitsch’s (2003) constructional representation, shown below. Let us 861
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 137
Sem BE-ABLE < >
Syn can SUBJ VPinf
IFor { }
IMy: Ability for
Action
Sem < agt act >
Syn can/could SUBJ2p VPinf
IFor request
Fig. 8.2 The Can you X? construction (Stefanowitsch 2003: 116)
first explain the assumptions behind this visualization. In (Cognitive) Construction 862
Grammar, constructions that share formal and functional aspects are linked to each 863
other in a (default) inheritance hierarchy (for details, see Goldberg 1995: Chapter 3). 864
Higher-level, schematic constructions pass on their properties to lower-level, more 865
concrete constructions, whose own specific properties may however block this 866
inheritance. In Fig. 8.2, the Can you X? construction, represented by the box at the 867
bottom, is linked to the more general Can construction expressing ability, shown 868
at the top, by a metonymy inheritance link (IMy): rather than making a request 869
directly, the Can you X? construction asks something that is closely associated with 870
requesting, namely whether the hearer is able to perform the requested action. This 871
inherent aspect of the request scenario is of course reminiscent of one of Searle’s 872
(1969) proposed felicity conditions on requesting. 873
In the Can you X? construction, unlike in the more general can construction, 874
the (IFor) parameter is filled in with the value ‘request’ but the Semantics (Sem) 875
parameter is left almost empty. While the semantics of the Can you X? construction 876
stipulates that there should be a thematic role of Agent present which is to 877
be expressed syntactically by the second person Subject argument and that the 878
complement of the modal should be an action, the modal verb itself does not carry 879
any semantic value. But even if the semantics of the modal verb is left empty 880
UNCORRECTED
PROOF
138 B. Cappelle
in this indirect speech act construction, the hearer might still involuntarily think 881
about ability via the construction’s link with the more general modal construction. 882
Nevertheless, Stefanowitsch adds between brackets: 883
I will leave open the question whether can is actually completely empty, or whether it retains 884
a weak trace of its meaning (Stefanowitsch 2003: 117) 885
In other words, the Can-you-X? construction contains pragmatic information 886
directly and semantic information either indirectly or in some strongly ‘diluted’ 887
form. 888
8.4.2 Short-Circuited Interpretation 889
Kay (2004) calls cases like Can you X? and the examples in (20) “strictly 890
interpretational constructions”, that is, constructions which “remain exclusively 891
interpretational, lacking idiosyncratic peculiarities of morphosyntax” (Kay 2004: 892
695).3 Cappelle and Depraetere (2014, 2016a) discuss a few further cases of these 893
interpretational constructions, still involving can (or could). I have listed them 894
in (21). Their conventionalized interpretation is revealed by the fact that it is 895
deliberately or otherwise ignored by the speaker, thus yielding a very marked, 896
and hence humoristic, use. The examples all come from various episodes of the 897
animation series The Simpsons. The relevant expressions are in italics. 898
899
900
(21) a. Lisa: I guess that’s it; these animals are all gonna die.Homer: Not if I can help it, Lisa.Lisa: Do you have an idea?Homer: Uh, no – sorry if it sounded like I did.
b. [context: Homer complains to Marge about his classmates bullying him when hewas a kid : : : ]Marge: Kids can be so cruel.Bart: We can? Thanks, Mom! [starts hitting Lisa]
c. Homer: Come on, boy. Let’s get me a six-pack.Man: Uh, sir, you can’t operate a boat under the influence of alcohol.Homer: That sounds like a wager to me!
d. Homer: Ladies and gentlemen, if I could just say a few words : : : I’d be a betterpublic speaker.
e. Marge [to Homer]: How’s your back, Homey?Homer: I can’t complain.[he then indicates a sign which reads: “No Complaining”]
901
3We pointed out above that morphosyntactically unpredictable properties do occur in the case ofCan you X?, namely the possible pre-verbal position that this construction shares with other directrequests. However, this element is not always present, and when it’s not, a Can you X? sentencelooks like any ordinary question.
, namely the possible pre-verbal position that this construction shares with other direct
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 139
f. Ned: Aw, gee, the man’s just trying to show his wife he cares for her. How can
we sabotage his labor of love?Homer: I dunno. Gasoline, acid, I got some stuff in the trunk.4
902
While in the general construction shown at the top of Fig. 8.2, can is specified as 903
having the meaning ‘BE-ABLE’, this modal can have a small set of basic meanings, 904
including not just ability but also permission, opportunity, and the like (Depraetere 905
and Reed 2011; Depraetere 2014). Not all of the clichéd expressions in (21) actually 906
have a short-circuited implicature. In some cases, all we can say about them is that 907
they come with a preselected choice for what the meaning of can is. And in some 908
cases, there is both a short-circuited semantic disambiguation and a short-circuited 909
implicature on top of that. For this reason, Cappelle and Depraetere (2014) extend 910
Morgan’s notion of short-circuited implicature to short-circuited interpretation, 911
which may hold at the level of semantics and at the level of pragmatics alike. 912
Thus, in (21a), Homer’s utterance uses what we could call the Not if I can 913
help it Construction, which specifies that can has as its meaning ‘HAVE-THE- 914
OPPORTUNITY’ and which moreover has the in-built implicature ‘I’ll prevent 915
this from happening’. The opportunity meaning doesn’t have to be bleached or 916
backgrounded at all by this pragmatic specification; they can sit comfortably 917
alongside each other. The semantic and pragmatic properties are present at the same 918
time in the interpretation of the whole (‘This won’t happen if I have the opportunity 919
to prevent it, and rest assured, I will take that opportunity’). Note also that the 920
form of this construction is rather fixed. While there are related, and similarly 921
conventional items (e.g. Not if I have anything to say about it), we cannot reverse 922
the order of the main clause fragment and the if -clause (*If I can help it, (then) not) 923
and a fuller version like This won’t happen if I can help it does not have the same 924
usage frequency and certainly does not feel like a stored item the way Not if I can 925
help it does. 926
The italicised expression in (21b) has a few hundred thousands of hits on Google. 927
There’s no doubt that adult speakers of English have heard this sequence several 928
times in their lifetime. One might be disinclined to call this a construction, since 929
4A similar example occurs in an episode of the British sitcom Not Going Out:
(i) Lee Anyway, er, it got me thinking. Why don’t we inviteyou two guys round for dinner?
Toby [posh upstairs neighbour,reluctant to accept]
Oh : : : That’s very kind.
Lee Great. Well, I will see you tonight at seven o’clock.Toby It’s very short notice to arrange a sitter.Lee Well, I’ll let you bring the baby round, eh? The more,
the merrier.Toby Right.Lee So how can you refuse?
Toby I’m not sure, give me a minute.
UNCORRECTED
PROOF
140 B. Cappelle
there is nothing very special about its form or interpretation. But note that an 930
existential reading (‘There are (many) kids which are cruel’ and/or ‘There are 931
(many) times at which kids are cruel’), which is merely compatible with the 932
meaning ‘GENERAL SITUATION POSSIBILITY’ (Depraetere and Reed 2011), 933
is an integral aspect of the standard interpretation of this sentence. Of course, Bart’s 934
elliptic reply We can? is cleverly used as a ‘hinge’, as it can innocuously refer back 935
to Marge’s utterance while at the same time functioning as if it is a standard indirect 936
speech act to express one’s exhilaration about being given permission for an action. 937
Why should we speak of a construction in (21c), the last example of this set that 938
we will look at in more detail here? This is not so easy to motivate, but to begin 939
with, the Uh, fVocative/Interjection(�like) expressiong part is a conventional way 940
to introduce something unpleasant for the hearer. While this does not necessarily 941
set the semantic value for the negated modal in the following part (you can’t X) 942
as ‘BE-PROHIBITED’, we might wonder whether Homer’s reply would sound 943
as ridiculous if this interpretation wasn’t short-circuited. Sure enough, Homer’s 944
next utterance only sounds funny because the prohibition interpretation is ignored 945
in favour of an ability reading of the sentence with can. It is unlikely that the 946
obviousness of a prohibition interpretation is purely the result of a pragmatic 947
inference process happening in context. Speakers may have heard similar cases and 948
have stored the relevant overlapping portion so that the interpretation becomes semi- 949
automatic. Web-attested examples include Uh, sir, you can’t park here, Uh, miss, 950
you can’t get on the bus with that umbrella open like that, Uh, hey, you can’t ask me 951
about my religious beliefs and Uh, excuse me, you can’t actually do that. 952
For more detailed overviews of how modality can be studied in Construction 953
Grammar, see Boogaart and Fortuin (2016), as well as the articles in Cappelle and 954
Depraetere (2016b). 955
8.4.3 Further Pragmatic Associations 956
Schmid (2014) suggests that I love you (as opposed to, say, She loved him or I 957
like soccer) is a “quasi-fixed lexico-grammatical unit” (p. 239), in other words, a 958
conventional expression stored as a construction, which activates “associations to 959
typical users (lovers, fiction writers, figures in movies), typical situations (romantic 960
moments), typical communicative intentions (assurance of deep affection)” (pp. 961
240–241). Schmid further argues that such pragmatic associations might even play 962
a role in the “chunking and freezing” (p. 241) of constructions, both relatively open 963
lexical bundles, such as I don’t know C wh-element, whereby “S informs H about 964
S’s lack of knowledge” (pp. 271–272) and complete sequences such as yes we 965
can, which as a special catchphrase has become part of the speech community’s 966
collective mind since Barack Obama’s 2008 election campaign. 967
The more people are familiar with an original formula, the more successfully 968
such a sequence can be used for deliberate modifications. For instance, yes we 969
can has spawned, even outside the community of native speakers of English, such 970
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 141
intentional variations as Yes, we can’t and No you can’t (Schmid 2014: 250, fn. 14), 971
as well as Yes We Candy!, Yes We CandlesTM , Yes We Carve! [Obama pumpkins], 972
Yes Wii Can, Yes We Can-Can Club, Yes we McCann, Yes We Cannes, Yes We 973
Canada, Yes We Canberra!, Yes We Cam’ [DCameron], Yes we Cat! [DCatalonia], 974
Yes We Con!, Yes we Khan, Yes We Plan, and so on, and so forth. Some linguists 975
have come to use the term ‘snowclone’ to refer to any “multi-use, customizable, 976
instantly recognizable, time-worn, quoted or misquoted phrase or sentence that can 977
be used in an entirely open array of different jokey variants by lazy journalists and 978
writers Pullum (2004).” The term is coined with specific reference to the rhetorical 979
trope If Eskimos have N [Dany number] words for snow, X surely have M words for 980
Y, based on the misguided cliché that Eskimos have innumerable distinct words for 981
snow (cf. Pullum 1991 for a debunking of this myth).5 982
Rasulic (2010) provides a Construction Grammar account of this sort of phe- 983
nomenon, discussing variations of aphoristic sequences such as once bitten, twice 984
shy and easy come, easy go, as well as changes of popular quotes such as And now 985
for something completely different (Monty Python’s Flying Circus) and Frankly my 986
dear, I don’t give a damn, spoken by Clark Gable in Gone with the Wind and voted 987
by a 1,500-member jury as the most memorable American movie quote of all time. 988
The twistings listed by Rasulic include Once married, twice shy, easy come, easy 989
blow, And now for something completely Madonna, and Frankly, my dear, I don’t 990
give a Daimler. Gonzálvez-García (2015) also deals with the linguistic properties 991
of snowclones from a constructionist and, moreover, contrastive (English-Spanish) 992
point of view, focussing mainly on the English saying A bird in the hand is worth 993
two in the bush and its Spanish counterpart Más vale pájaro en mano que ciento 994
volando (lit. ‘Better a bird in the hand than a hundred flying’) and their various 995
manipulations, such as A beer in the hand is worth six in the fridge and Más vale 996
dinero en mano que bonos volando (‘money in the hand is better than bonuses in 997
the air’). Finally, Fiorentini and Marino (2015) discuss the formula “KEEP CALM 998
AND X”, based on the well-known and visually iconic English propaganda poster, 999
from a semiotic point of view. As the authors point out, this pattern “(as well as its 1000
counterparts in other languages, such as Italian) is a widespread formula used over 1001
the Web, in advertising, and in everyday interactions as a softened imperative form, 1002
which rapidly developed new pragmatic meanings.” 1003
As with the comical misuses of modal verb constructions in The Simpsons 1004
mentioned in Sect. 8.4.2, adapting a formulaic expression can only have a humorous 1005
effect if the addressee is familiar with the ‘correct’ (i.e. standard, conventionalized) 1006
expression. Snowclones thus demonstrate that we assume our addressees have 1007
stored in their long-term memory a vast number of (supposedly) fixed expressions, 1008
many of which have their origin in popular culture. This shared familiarity with the 1009
larger cultural context and other so-called encyclopaedic knowledge (knowledge 1010
about ‘the world’) also ties in with what Frame Semantics endeavours to capture 1011
5A large collection of snowclones in English is maintained by O’Connor (2007). Cappelle (2013)lists a few snowclones in Dutch.
writers Pullum (2004).” The term is coined with specific reference to the rhetorical
UNCORRECTED
PROOF
142 B. Cappelle
(see, e.g., Boas 2005, Fillmore 1976, Fillmore and Atkins 1992, Petruck 1996). 1012
Interestingly, note that this sister theory of Construction Grammar is called Frame 1013
Semantics and not Frame Pragmatics, a point raised by Andor (2010) in his 1014
interview with Charles Fillmore. It should be clear, though, that allusions to well- 1015
known standard formulae or to famous quotes etc. have a different status from the 1016
contents proper of a sentence. Suppose your interlocutor says “If lusting after naked 1017
Daniel Radcliffe is wrong, I don’t want to be right” and you then say “Amen to 1018
that”, what you agree with is that Daniel Radcliffe is an object of desire; you don’t 1019
agree with the contents of the presumed original from a 1970s soul song: “If loving 1020
you is wrong, I don’t want to be right”, even though you might have appreciated the 1021
creative exploitation of such a stock phrase. That is why it makes sense to distinguish 1022
purely semantic and less direct aspects of communication. 1023
8.4.4 And Then There’s the Reminder Existential Construction 1024
Lakoff (1987) discusses the ‘reminder’ use of the existential construction, illustrated 1025
in (22): 1026
1027
1028
(22) a. There’s always Harry.b. But there’s the dog!c. There’s still the remains of Christmas dinner in the freezer. (Lakoff 1987: 561)
1029
Now, consider how Lakoff comments on this construction – or on a conventional 1030
use of it: 1031
Examples of reminders are most commonly given in the form of lists: 1032
– There’s the cat to feed, the dog to walk, the horse to brush, : : : 1033
– There’ll be Max at the head of the table, Sally next to me, : : : 1034
Lists of this sort are a general feature of English and are not peculiar to reminding uses of 1035
existentials. They apply to all sorts of other constructions: 1036
– Joan is prettier than Sue, richer than Melanie, smarter than Eliza, : : : 1037
– Bring the camera, the backpack, the canteen, : : : 1038
– I want to give Tom a sweater, Jeff an espresso-maker, : : : 1039
– Tom likes cats, Sally horses, Mike dogs, : : : 1040
Existential lists are simply cases where the list construction has applied to an existential 1041
sentence. They are not part of the analysis of existentials at all. (Lakoff 1987: 561–562) 1042
What Lakoff is suggesting here is that the use of a list does not have to be stored as 1043
specific information in the Reminder Existential Construction. It concerns a general 1044
fact of language. But let us consider the following authentic example: 1045
1046
1047
(23) The smartphone is typical of Amazon. There is the remorseless expansion: if you candeliver books and washing machines, why not a phone? There is the ability to switchbetween the real world of atoms and the digital world of bits ( : : : ). There is the drivefor market share over immediate profits. And there is the slightly creepy feeling thatAmazon knows too much about its users already. (The Economist, 21 June 2014, p. 9)
1048
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 143
It seems to me that if there was just one occurrence of There is the : : : in the above 1049
example, this would be strange. The first sentence with a Reminder Existential 1050
Construction raises the expectation that some more such sentences are about to 1051
come. The typical use of this construction in a list is a specification that I would 1052
argue must be stored. How else would Lakoff (1987), in the cited passage above, 1053
be able to assert that examples of this construction are most typically provided in 1054
list form? This list use, we might say, is an implication of the Reminder Existential. 1055
(An implication is not be confused with an implicature, which is a special kind of 1056
implication.) As with any implication, language users pay attention to it and update 1057
their knowledge of the pattern with each new occurrence of it. As Bybee (2006) 1058
writes: 1059
It appears that listeners and speakers follow and keep track of the implications that occur 1060
in particular contexts. In order to know that a certain implication has occurred frequently 1061
and is associated with a certain string of words, speakers must register the context and the 1062
implications from the very first exposure. (Bybee 2006: 22) 1063
This means that there cannot be any sharp distinction between stored aspects 1064
of language and computed aspects of language (cf. Sect. 8.3). Even if aspects 1065
of an expression’s interpretation or use in discourse may initially be processed 1066
compositionally by a learner, these aspects will necessarily also be stored away 1067
in long-term memory from these first encounters. This continued storage will go 1068
on until the non-purely semantic aspects of a pattern are sufficiently strongly 1069
represented in the mind not to be derived anew, in each subsequent usage event, 1070
from knowledge about other aspects of the grammar. 1071
In the case of the Reminder Existential Construction, where in the construction’s 1072
representation should we provide place for the more or less strong intuition that 1073
this construction is especially common in a list context? While a definitive answer 1074
awaits more careful consideration, it is clear for now that this kind of information 1075
does not belong in the Semantics (Sem) part, which Construction Grammarians like 1076
to reserve for propositional semantics (cf. Fillmore et al. 1988), thematic roles, or 1077
the basic, core meaning of an item (e.g. ‘POSSIBILITY’ for can). Nor does it belong 1078
in the Illocutionary force (IFor), which as we saw concerns indirect speech act 1079
information. We might therefore need a separate section, used for other information 1080
about how a construction is typically used, perhaps together with information- 1081
packing properties, connotative value, register and region specifics etc. 1082
8.5 Concluding Reflections 1083
In this chapter, we have taken a look at various pragmatic phenomena – Gricean 1084
maxims, information structure, indirect speech acts, etc. – and the treatment they 1085
have received in Construction Grammar. We have shown how cross-linguistic or 1086
even language-internal evidence can force us to conclude that pragmatic information 1087
is conventionalized and therefore has to be learned and stored. In view of the 1088
topic of this volume – the semantics-pragmatics interface – let us finally reflect 1089
UNCORRECTED
PROOF
144 B. Cappelle
on the distinction, if any, that people make between these two aspects of linguistic 1090
knowledge. 1091
From its early days onwards, Construction Grammar has represented construc- 1092
tions as Saussurian signs, with just two poles: 1093
Grammatical constructions [ : : : ] are complex cognitive models with two dimensions: one 1094
characterizing parameters of form and one characterizing parameters of meaning. (Lakoff 1095
1987: 482) 1096
Construction Grammar thus seems to be a theory about bipolar units in this 1097
view. Even if Fillmore et al. (1988) recognized a separate layer of pragmatics, 1098
such a tripartite structure of constructions as containing syntactic, semantic and 1099
pragmatic information has often been reduced to a simpler structure. Fillmore 1100
(2013: 112) speaks of grammatical constructions as “the rules that unite formal 1101
and semantic information into various kinds of linguistic objects”, clarifying in a 1102
footnote that “‘[f]ormal’ here includes syntactic, morphological, and phonological 1103
form; ‘semantic’ includes pragmatics and conventions of usage” (Fillmore 2013: 1104
131). Yet, just because constructions are represented as binary in form, this need not 1105
mean each half is undivided and that semantics actually subsumes pragmatics: 1106
Langacker defines a grammar as a structured inventory of conventional linguistic units. 1107
The conventional linguistic units are symbolic units, and their two halves, form and 1108
meaning. Cognitive Grammar emphasizes the symbolic character of the linguistic sign 1109
(to use the Saussurean term). Langacker argues that the properties of constructions, as 1110
broadly defined, fall into two categories, which we describe here as form (the signifier) 1111
and meaning or function (the signified): the formal properties are syntactic, morphological, 1112
and phonological, and the functional properties are semantic, pragmatic, and discourse- 1113
functional. A construction is thus a symbolic unit, linking form and function as a symbol or 1114
sign. (Croft 2007: 490; emphasis mine) 1115
A question that will have to be more deeply explored in further research is to which 1116
extent we need to make a neat distinction within the functional pole of constructions 1117
between semantic and other information. Goldberg (1995) is rather equivocal about 1118
this when she writes that “[a] notion rejected by Construction Grammar is that 1119
of a strict division between semantics and pragmatics. Information about focused 1120
constituents, topicality, and register is presented in constructions alongside semantic 1121
information” (Goldberg 1995: 7). So, on the one hand, Goldberg here says that 1122
a strict semantics/pragmatics distinction is eschewed but on the other, the list of 1123
pragmatic kinds of information is treated as complementing semantic information 1124
rather than merging with it to form an undifferentiated bag of functional aspects. In 1125
this chapter I have offered a few pointers at how this question may find an answer. 1126
First, we consider the notion of short-circuited implicature useful and have 1127
briefly demonstrated how it can be extended to short-circuited interpretation more 1128
generally. Both semantics and illocutionary force are needed in some constructions. 1129
For instance, in the sequence If I could just say a few words, the semantics 1130
of the modal verb is that of ‘HAVE-PERMISSION’ (rather than ‘BE-ABLE’ or 1131
‘HAVE-OPPORTUNITY’), so this short-circuited interpretational aspect of this 1132
sequence is of the semantic kind. The conventional use of this expression is a short- 1133
circuited interpretational aspect of the pragmatic kind. This use includes the hearer’s 1134
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 145
expectation that the speaker will not wait for an actual response to see whether or 1135
not the permission has been granted but uses this expression as an accepted hedge 1136
to make an announcement and to ask thereby for some silence from the audience 1137
(cf. Declerck and Reed 2001: 356). 1138
Second, there is other ‘pragmatic’ information than indirect speech act infor- 1139
mation that is stored in constructions. This information may describe everything 1140
we need to know about the larger context in which the construction can be 1141
embedded and the speech situation in which it can be used. Here we find stored 1142
information-structural constraints, information about how the construction helps 1143
structure discourse, as well as register and politeness values (e.g. constraints relating 1144
to social relationships) and (if relevant) allusions to the broader world of literature, 1145
popular culture, etc. All this information might be grouped together under a single 1146
label, although it remains to be seen whether information-structural properties might 1147
not need a separate treatment. This may well turn out to be the case, as being topical, 1148
focused, accessible, etc. usually applies to parts of a construction and not to the 1149
construction as a whole. We thus propose the provisional representational structure 1150
of sentence-level constructions shown in Fig. 8.3. 1151
One thing that has become evident, I hope, is that there is a lot of conventional 1152
knowledge involved in how we understand a sentential construction. Some of this 1153
understanding, of course, is semantic in nature, dealing with the propositional 1154
content of the utterance. Apart from semantic information, we also make use of 1155
pragmatic information in interpreting a construction in use, but not everything that 1156
is pragmatic about this interpretation is necessarily to be considered unpredictably 1157
context-dependent. There is much pragmatics that is conventionally linked to 1158
constructions. Semantics and pragmatics can live peacefully side by side in a single 1159
construction. 1160
Sem
Syn
Prag IFor
Information
Structure
Other Discourse
organisation
Register
Fig. 8.3 A template of constructional representations
UNCORRECTED
PROOF
146 B. Cappelle
Acknowledgements I am grateful to Ilse Depraetere for her valuable input to the Simpsons 1161
examples discussed in Sect. 8.4. I benefited from discussions with Yukio Hirose, Masaru Kanetani 1162
and Naoaki Wada ensuing my presentation of a first sketchy version of this text at the University 1163
of Tsukuba. As I drafted this version in a room of the University of Tsukuba’s Guest House in 1164
which Charles Fillmore had previously stayed, I would like to believe I have been particularly 1165
inspired by some of his lasting legacy. Many thanks also to Friedemann Pulvermüller for giving 1166
me the opportunity to present the main points of this text at the Freie Universität Berlin in early 1167
2015; I also thank the members of the audience there for their stimulating questions. Finally, I 1168
would like to thank Ilse Depraetere and Raphael Salkie for their patience with the slow progress of 1169
this contribution, and an anonymous reviewer for some constructive comments. Though the cited 1170
literature in this chapter was not meant to be exhaustive or even representative, I alone remain 1171
responsible for any major oversights or misrepresentations, as well as for any other shortcomings 1172
of my text. 1173
References 1174
Aijmer, K. (1997). I think – An English modal particle. In T. Swan & O. J. Westvik (Eds.), Modality 1175
in Germanic languages. Historical and comparative perspectives (pp. 1–47). Berlin: Mouton 1176
de Gruyter. 1177
Akmajian, A. (1984). Sentence types and the form-function fit. Natural Language and Linguistic 1178
Theory, 2(1), 1–23. 1179
Ambridge, B., & Goldberg, A. E. (2008). The island status of clausal complements: Evidence in 1180
favor of an information structure explanation. Cognitive Linguistics, 19, 357–389. 1181
Andor, J. (2010). Discussing frame semantics: The state of the art. An interview with Charles 1182
J. Fillmore. Review of Cognitive Linguistics, 8(1), 157–176. 1183
Bergs, A. (2008). Can we take Construction Grammar beyond sneezing napkins off tables? In 1184
K. Stierstorfer (Ed.), Proceedings of the Anglistentag (Münster 2007, pp. 269–276). Trier: 1185
WVT. 1186
Bergs, A., & Diewald, G. (2009). Contexts and constructions. Amsterdam: John Benjamins. 1187
Boas, H. C. (2005). From theory to practice: Frame semantics and the design of FrameNet. In 1188
S. Langer & D. Schnorbusch (Eds.), Semantik im Lexikon (pp. 129–160). Tübingen: Narr. 1189
Boas, H. C. (Ed.). (2010). Contrastive studies in Construction Grammar. Amsterdam: John 1190
Benjamins. 1191
Bolinger, D. (1971). The phrasal verb in English. Cambridge, MA: Harvard University Press. 1192
Boogaart, R. (2015). ‘Toen ik je zag’: de kracht van losse bijzinnen. [‘When I saw you’: the power 1193
of free subclauses’]. Onze Taal, 84(1), 12–14. 1194
Boogaart, R., & Fortuin, E. (2016). Modality and mood in cognitive linguistics and construction 1195
grammars. In J. van der Auwera & J. Nuyts (Eds.), The Oxford handbook of mood and modality 1196
(pp. 514–533). Oxford: Oxford University Press. 1197
Booij, G. (2010). Construction morphology. Oxford: Oxford University Press. 1198
Booij, G., & Hüning, M. (2014). Affixoids and constructional idioms. In R. Boogaart, T. Colleman, 1199
& G. Rutten (Eds.), Extending the scope of Construction Grammar (pp. 77–105). Berlin: 1200
Mouton de Gruyter. 1201
Brinton, L. J. (2011). The extremes of insubordination: Exclamatory as if! Paper presented at the 1202
second conference of the International Society of the Linguistics of English (ISLE 2), Boston, 1203
17–21 June 2011. 1204
Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language use. Cambridge: 1205
Cambridge University Press. 1206
Bybee, J. (2006). From usage to grammar: The mind’s response to repetition. Language, 82(4), 1207
711–733. 1208
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 147
Bybee, J. (2013). Usage-based theory and exemplar representation of constructions. In Thomas 1209
Hoffmann & Graeme Trousdale (Eds.) (pp. 49–69). 1210
Cappelle, B. (2003, May 13). Het gebruik van het Brugse dialoogpartikel dè (/dæ/). Neerlan- 1211
distiek.nl. http://www.neerlandistiek.nl/03.01/ 1212
Cappelle, B. (2005). Particle patterns in English: A comprehensive coverage. Ph.D. dissertation. 1213
University of Leuven. 1214
Cappelle, B. (2009). Contextual cues to particle placement: Multiplicity, motivation, modeling. In 1215
Bergs and Diewald (Eds.) (pp. 145–191). 1216
Cappelle, B. (2013). De sneeuwkloon: kroniek van een aangekondigde dooddoener. Over Taal, 1217
52(1), 6–8. 1218
Cappelle, B. (2014). Conventional combinations in pockets of productivity: English resultatives 1219
and Dutch ditransitives expressing excess. In R. Boogaart, T. Colleman, & G. Rutten (Eds.), 1220
Extending the scope of Construction Grammar (pp. 251–282). Berlin: Mouton de Gruyter. 1221
Cappelle, B., & Depraetere, I. (2014). Implicated modal meaning in Construction Grammar. 1222
Paper presented at the workshop on “Modal meaning in Construction Grammar” held at the 1223
Conference of the International Society for the Linguistics of English (ISLE), 24–27 August 1224
2014. 1225
Cappelle, B., & Depraetere, I. (2016a). Short-circuited interpretations of modal verb constructions: 1226
Some evidence from The Simpsons. Constructions and Frames, 8(1), 7–39. 1227
Cappelle, B., & Depraetere, I. (eds.). (2016b). Special issue on modal meaning in Construction 1228
Grammar. Constructions and Frames, 8(1). 1229
Cappelle, B., Dugas, E., & Tobin, V. (2015). An afterthought on let alone. Journal of Pragmatics, 1230
80, 70–85. 1231
Claridge, C., & Arnovick, L. (2010). Pragmaticalisation and discursisation. In A. H. Jucker & 1232
I. Taavitsainen (Eds.), Historical pragmatics (pp. 165–192). Berlin/New York: Mouton de 1233
Gruyter. 1234
Coussé, E., & Oosterhof, A. (2012). Het imperativische participium in het Nederlands: vorm, 1235
betekenis en gebruik. Nederlandse Taalkunde, 17(1), 26–56. 1236
Croft, W. (1998). Linguistic evidence and mental representations. Cognitive Linguistics, 9, 1237
151–173. 1238
Croft, W. (2007). Construction Grammar. In D. Geeraerts & H. Cuyckens (Eds.), The Oxford 1239
handbook of cognitive linguistics (pp. 463–508). Oxford: Oxford University Press. 1240
D’Hertefelt, S., & Jean-Christophe, V. (2014). Independent complement constructions in Swedish 1241
and Danish: Insubordination or dependency shift? Journal of Pragmatics, 60, 89–102. 1242
Declerck, R., & Reed, S. (2001). Conditionals: A comprehensive empirical analysis. Mouton: De 1243
Gruyter. 1244
Del Campo Martínez, N. (2013). Illocutionary constructions in English: Cognitive motivation and 1245
linguistic realization. A study of the syntactic realizations of the directive, commissive and 1246
expressive speech acts in English. Bern: Peter Lang. 1247
Del Campo Martínez, N., & de Mendoza, F. J. R. (2012). A constructionist approach to illocution: 1248
The case of orders. Miscelánea Journal of English and American Studies, 45, 13–31. 1249
Depraetere, I. (2014). Modals and lexically-regulated saturation. Journal of Pragmatics, 71, 1250
160–177. 1251
Depraetere, I., & Reed, S. (2011). Towards a more explicit taxonomy of root possibility. English 1252
Language and Linguistics, 15(1), 1–29. 1253
Deulofeu, H.-J., & Debaisieux, J.-M. (2009). Constructions and context: When a construction 1254
constructs the context. In Bergs and Diewald (Eds.), (pp. 43–62). 1255
Diewald, G. (2011). Pragmaticalization (defined) as grammaticalization of discourse functions. 1256
Linguistics, 49, 365–390. 1257
Ebensgaard, K. (2014, June 4). Review of Del Campo Martínez (2013). LINGUIST List, 1258
Vol 25–2425.http://linguistlist.org/issues/25/25-2425.html 1259
Erman, B., & Kotsinas, U.-B. (1993). Pragmaticalization: The case of ba’ and you know. Studier i 1260
Modernspråkvetenskap, 10, 76–93. 1261
UNCORRECTED
PROOF
148 B. Cappelle
Espinal, M. T., & Mateu, J. (2010). On classes of idioms and their interpretation. Journal of 1262
Pragmatics, 42, 1397–1411. 1263
Fillmore, C. C. (1976). Frame semantics and the nature of language. In S. R. Harnad, H. D. Steklis, 1264
& L. Jane (Eds.), Annals of the NY Academy of Sciences (Origins and Evolution of Language 1265
and Speech, Vol. 280, pp. 20–32). New York: New York Academy of Sciences. 1266
Fillmore, C. J. (2013). Berkeley Construction Grammar. In Thomas Hoffmann and Graeme 1267
Trousdale (Eds.) (pp. 111–132). 1268
Fillmore, C. J., & Atkins, B. T. (1992). Towards a frame-based organization of the lexicon: The 1269
semantics of RISK and its neighbors. In A. Lehrer & E. Kittay (Eds.), Frames, fields and 1270
contrast: New essays in semantics and lexical organization (pp. 75–102). Hillsdale: Erlbaum. 1271
Fillmore, C. J., Kay, P., & O’Connor, M. C. (1988). Regularity and idiomaticity in grammatical 1272
constructions: the case of let alone. Language, 64, 501–538. 1273
Fiorentini, I., & Marino, G. (2015). New literacies and linguistic innovation: The case of Internet 1274
memes. Abstract for a paper presented at the 48th annual meeting of the Societas Linguistica 1275
Europaea. September 2–5, 2015. Leiden University Center for Linguistics (LUCL), The 1276
Netherlands. http://sle2015.eu/downloads/BOOK%20OF%20ABSTRACTS_2.pdf 1277
Fortin, A. (2011). The morphology and semantics of expressive affixes. Ph.D. dissertation. Oxford 1278
University. 1279
Fortuin, E. (2003). De directieve infinitief en de imperatief in het Nederlands. Nederlandse 1280
taalkunde, 8, 14–43. 1281
Fried, M., & Östman, J.-O. (2005). Construction grammar and spoken language: The case of 1282
pragmatic particles. Journal of Pragmatics, 37(11), 1752–1778. 1283
Glasbey, S. R. (2003). Let’s paint the town red for a few hours: Composition of aspect in idioms. In 1284
A. M. Wellington (Ed.), Proceedings of the ACL workshop: The lexicon and figurative language 1285
(pp. 42–48). Sapporo, Japan. 1286
Goldberg, A. E. (1995). Constructions: A Construction Grammar approach to argument structure. 1287
Chicago: University of Chicago Press. 1288
Goldberg, A. E. (2000). Patient arguments of causative verbs can be omitted: The role of 1289
information structure in argument distribution. Language Sciences, 34, 503–524. 1290
Goldberg, A. E. (2003). Constructions: A new theoretical approach to language. Trends in 1291
Cognitive Sciences, 7(5), 219–224. 1292
Goldberg, A. E. (2004). Argument realization: The role of constructions, lexical semantics and 1293
discourse factors. In J.-O. Östman & M. Fried (Eds.), Construction Grammars: Cognitive 1294
grounding and theoretical extensions (pp. 17–44). Amsterdam: John Benjamins. 1295
Gonzálvez-García, F. (2015). On the gradient nature of constructions: Evidence from similes 1296
and proverbs in English and Spanish. Paper presented at the 2nd international symposium on 1297
Figurative Thought and Language (FTL2), October 28–30. University of Pavia, Italy. 1298
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics 1299
III: Speech acts (pp. 41–58). New York: Academic. 1300
Gries, S. T. (2003). Multifactorial analysis in corpus linguistics: A study of particle placement. 1301
London: Continuum. 1302
Haegeman, L. (1993). The interpretation of the particle da in West Flemish. Lingua, 90(1/2), 1303
111–128. 1304
Hampe, B. (2002). Superlative verbs: A corpus-based study of semantic redundancy in English 1305
verb-particle constructions. Tübingen: Gunter Narr. 1306
Hilpert, M. (2014). Construction Grammar and its application to English. Edinburgh: Edinburgh 1307
University Press. 1308
Hirose, Y. (1992). Because no metagengoteki kino nitsuite [On a metalinguistic use of because], 1309
Eigo Onseigaku to Eigo Kyoiku [English Phonetics and English Education] (pp. 81–85). Tokyo: 1310
Kairyudo. 1311
Hoffmann, T., & Bergs, A. (2012). ‘Are you construction in disguise : : : ’: Investigating the role 1312
of context in football chant constructions. Paper presented at the 7th international conference 1313
on construction grammar. August 9–12, 2012. Hankuk University of Foreign Studies, Seoul, 1314
South Korea. 1315
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 149
Hoffmann, T., & Trousdale, G. (Eds.). (2013). The Oxford handbook of Construction Gammar. 1316
Oxford: Oxford University Press. 1317
Jackendoff, R. (2002). Foundations of language: Brain, meaning, grammar, evolution. Oxford: 1318
Oxford University Press. 1319
Jackendoff, R. (2010). Meaning and the lexicon: The parallel architecture 1975–2010. Oxford: 1320
Oxford University Press. 1321
Jurafsky, D. (1996). Universal tendencies in the semantics of diminutives. Language, 72(3), 1–47. 1322
Kaltenböck, G. (2006). On the grammatical structure of insubordinate if-clauses. In G. Kaltenböck, 1323
E. Keizer, & A. Lohmann (Eds.), Outside the clause: Form and function of extra-clausal 1324
constituents (pp. 341–378). Amsterdam: John Benjamins. 1325
Kanetani, M. (2009). Metalinguistic reason constructions revisited. Tsukuba English Studies, 27, 1326
27–44. 1327
Kanetani, M. (2012). Another look at the metalinguistic Because-Clause Construction. Tsukuba 1328
English Studies, 31, 1–18. 1329
Kay, P. (2004). Pragmatic aspects of grammatical constructions. In L. Horn & G. Ward (Eds.), The 1330
handbook of pragmatics (pp. 675–700). London: Blackwell. 1331
Kay, P., & Fillmore, C. (1999). Grammatical constructions and linguistic generalizations: The 1332
what’s X doing Y? construction. Language, 75, 1–33. 1333
Kudo, S. (2011). A comparative study between Resultative Constructions and Body Part Off 1334
Constructions. Tsukuba English Studies, 29, 169–185. 1335
Lakoff, G. (1987). Women, fire, and dangerous things: What categories reveal about the mind. 1336
Chicago: The University of Chicago Press. 1337
Lambrecht, K. (1990). ‘What me worry?’ Mad magazine sentences revisited. In K. Hall et al. 1338
(Eds.), Proceedings of BLS 16 (pp. 215–228). Berkeley: Berkeley Linguistics Society. 1339
Lambrecht, K. (1994). Information structure and sentence form: Topic, focus and the mental 1340
representations of discourse referents. Cambridge: Cambridge University Press. 1341
Lambrecht, K. (1995). Compositional vs. constructional meaning: The case of French comme-N. 1342
In S. Mandy & G. Teresa (Eds.), Proceedings of SALT 5 (186–203). Ithaca: Comell University. 1343
Lambrecht, K. (2004). On the interaction of information structure and formal structure in 1344
constructions: The case of French Right-detached comme-N. In M. Fried & J.-O. Östman 1345
(Eds.), Construction Grammar in a cross-language perspective (pp. 157–199). Amsterdam: 1346
John Benjamins. 1347
Langacker, R. (1987). Foundations of Cognitive Grammar (Vol. 1). Stanford: Stanford University 1348
Press. 1349
Lauwers, P., Vanderbauwhede, G., & Verleyen, S. (Eds.). (2012). Pragmatic markers and 1350
pragmaticalization: Lessons from false friends. Amsterdam: John Benjamins. 1351
Leech, G. (2014). The pragmatics of politeness. Oxford: Oxford University Press. 1352
Lee-Goldman, R. (2011). Context in constructions. Ph.D. dissertation. University of California at 1353
Berkeley. 1354
Leino, J. (2013). Information Structure. In Thomas Hoffmann and Graeme Trousdale (Eds.) 1355
(pp. 329–344). 1356
Lemmens, M. (2006). More on objectless transitives and ergativization patterns in English. 1357
Constructions 1. urn:nbn:de:0009-4-6802. 1358
Mateu, J., & Espinal, T. M. (2007). Laughing our heads off: When metaphor constrains aspect. 1359
Handout for a paper presented at the 33rd annual meeting of the Berkeley Linguistics Society. 1360
February 9–11, 2007. University of California, Berkeley. 1361
Michaelis, L. A., & Lambrecht, K. (1996). Toward a construction-based model of language 1362
function: The case of nominal extraposition. Language, 72, 215–247. 1363
Morgan, J. L. (1977). Two types of convention in indirect speech-acts. Technical Report No. 52. 1364
University of Illinois at Urbana-Champaign. 1365
Morris, C. (2013). Bekje-af, knettertjegek en dolletjesgelukkig: The use and development of 1366
intensifying diminutive compound in Dutch within the framework of Construction Morphology. 1367
MA thesis. University of Groningen. 1368
Kaltenböck, G. (2006). On the grammatical structure of insubordinate if-clauses. In G. Kaltenböck,
UNCORRECTED
PROOF
150 B. Cappelle
Morris, C. (2014, September 13). “Ik schrik me : : : ”. Pelikanenschurft. http:// 1369
pelikanenschurft.wordpress.com/2014/09/13/ik-schrik-me/, last Accessed 9 Nov 2014. 1370
Nikiforidou, K. (2009). Constructional analysis. In F. Brisard, J.-O. Östman, & J. Verschueren 1371
(Eds.), Grammar, meaning and pragmatics (pp. 16–32). Amsterdam: John Benjamins. 1372
Norde, M., & Morris, C. (2014). Category change and constructional schemas: The case of 1373
diminutive intensifying prefixoids in Dutch. Paper presented at the 8th international conference 1374
on construction grammar. September 3–6, 2014. University of Osnabrück, Germany. 1375
O’Connor, E. (2007). The snowclones database. http://snowclones.org/ 1376
Östman, J.-O. (1999). Coherence through understanding through discourse patterns: Focus on news 1377
reports. In W. Bublitz, U. Lenk, & E. Ventola (Eds.), Coherence in spoken and written discourse 1378
(pp. 77–100). Amsterdam: John Benjamins. 1379
Östman, J.-O. (2000). Postkortsdiskurs: med den språkliga periferin som centrum [Postcard 1380
discourse: placing the linguistic periphery at the center]. Sphinx 1999–2000 [The Yearbook 1381
of the Finnish Society of Sciences and Letters], 7–26. 1382
Östman, J.-O. (2005). Construction discourse: A prolegomenon. In J.-O. Östman & M. Fried 1383
(Eds.), Construction grammars: Cognitive grounding and theoretical extensions (pp. 121–144). 1384
Amsterdam: John Benjamins. 1385
Petruck, M. R. L. (1996). Frame semantics. In J. Verschueren, J.-O. Ostman, J. Blommaert, & C. 1386
Bulcaen (Eds.), Handbook of pragmatics (pp. 1–13). Philadelphia: John Benjamins. 1387
Potts, C. (2007). The expressive dimension. Theoretical Linguistics, 33(2), 165–197. 1388
Pullum, G. K. (1991). The great Eskimo vocabulary hoax and other irreverent essays on the study 1389
of language. Chicago: The University of Chicago Press. 1390
Pullum, G. K. (2004, January 16). Lexicographical dating to the second. Language Log. http:// 1391
itre.cis.upenn.edu/~myl/languagelog/archives/000350.html 1392
Pullum, G. K. (2013). X, let alone Y. In Language Log. http://languagelog.ldc.upenn.edu/nll/ 1393
?p=4553. Accessed 2 Feb 2014. 1394
Rasulic, K. (2010). Long time, no buzz: Fixed expressions as constructional frames. CogniTextes 5. 1395
http://cognitextes.revues.org/356 1396
Ross, J. R.. (1967). Constraints on variables in syntax. PhD dissertation. Massachusetts Institute 1397
of Technology. 1398
Sadock, J. M. (1974). Toward a linguistic theory of speech acts. New York: Academic Press. 1399
Salmon, W. (2015). Conversational implicatures, reference point constructions, and that noun 1400
thing. Linguistics, 53(3), 443–477. 1401
Schmid, H.-J. (2014). Lexico-grammatical patterns, pragmatic associations and discourse fre- 1402
quency. In T. Herbst, H.-J. Schmid, & S. Faulhaber (Eds.), Constructions, collocations, patterns 1403
(pp. 239–293). Berlin: Mouton de Gruyter. 1404
Schwenter, S. A. (1996). The pragmatics of independent si-clauses in Spanish. Hispanic Linguis- 1405
tics, 8, 316–351. 1406
Schwenter, S. A. (2016). Independent si-clauses in Spanish: Functions and consequences for 1407
insubordination. In N. Evans & H. Watanabe (Eds.), Insubordination (pp. 89–112). Amsterdam: 1408
John Benjamins. 1409
Searle, J. R. (1969). Speech acts. An essay in the philosophy of language. Cambridge: Cambridge 1410
University Press. 1411
Stefanowitsch, A. (2003). A construction-based approach to indirect speech acts. In K.-U. Panther 1412
& L. Thornburg (Eds.), Metonymy and pragmatic inferencing (pp. 105–126). Amsterdam: John 1413
Benjamins. 1414
Szczesniak, K., & Pachoł, M. (2015). What? Me, lie? The form and reading of the Incredulity 1415
Response Construction. Constructions online. http://www.constructions-journal.com; ISSN: 1416
1860–2010. 1417
Trousdale, G. (2014). Variation and construction grammar: The case of English hypocoristics. 1418
Plenary lecture at the conference of the International Society for the Linguistics of English 1419
(ISLE), 24–27 August 2014. 1420
Van Goethem, K., & Hiligsmann, P. (2014). When two paths converge: Debonding and clipping of 1421
Dutch reuze. Journal of Germanic Linguistics, 26(1), 31–64. 1422
UNCORRECTED
PROOF
8 What’s Pragmatics Doing Outside Constructions? 151
Van linden, A., & Van de Velde, F. (2014). (Semi-)autonomous subordination in Dutch: Structures 1423
and semantic-pragmatic values. Journal of Pragmatics, 60, 226–250. 1424
Van Olmen, D. (2009). De imperativische infinitief in het Nederlands: Een corpusgebaseerde 1425
benadering. Nederlandse Taalkunde, 14, 147–170. 1426
Van Olmen, D. (2010). Typology meets usage: The case of the prohibitive infinitive in Dutch. Folia 1427
Linguistica, 44(2), 471–507. 1428
Wide, C. (2009). Interactional Construction Grammar. Contextual features of determination 1429
in dialectal Swedish. In B. Alexander & D. Gabriele (Eds.), Contexts and construc- 1430
tions(pp. 111–141). Amsterdam: John Benjamins. 1431