The Collapse of Insensitive Semantics

See
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/227015639
The collapse of insensitive semantics
Article in Linguistics and Philosophy · April 2010

DOI: 10.1007/s10988-010-9077-y
CITATIONS READS
2 37
1 author:
Friedrich Christoph Doerge

University of Tuebingen
12 PUBLICATIONS 10 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Nature of (Linguistic) Meaning View project
'What is an "illocutionary act"?' View project
All content following this page was uploaded by Friedrich Christoph Doerge on 12 February 2014.
The user has requested enhancement of the downloaded file.

Dear Author,
Here are the proofs of your article.
• You can submit your corrections online, via e-mail or by fax.
• For online submission please insert your corrections in the online correction form. Always
indicate the line number to which the correction refers.
• You can also insert your corrections in the proof PDF and email the annotated PDF.
• For fax submission, please ensure that your corrections are clearly legible. Use a fine black
pen and write the correction in the margin, not too close to the edge of the page.
• Remember to note the journal title, article number, and your name when sending your
response via e-mail or fax.
• Check the metadata sheet to make sure that the header information, especially author names
and the corresponding affiliations are correctly shown.
• Check the questions that may have arisen during copy editing and insert your answers/
corrections.
• Check that the text is complete and that all figures, tables and their legends are included. Also
check the accuracy of special characters, equations, and electronic supplementary material if
applicable. If necessary refer to the Edited manuscript.
• The publication of inaccurate data such as dosages and units can have serious consequences.
Please take particular care that all such details are correct.
• Please do not make changes that involve only matters of style. We have generally introduced
forms that follow the journal’s style.
Substantial changes in content, e.g., new results, corrected values, title and authorship are not
allowed without the approval of the responsible editor. In such a case, please contact the
Editorial Office and return his/her consent together with the proof.
• If we do not receive your corrections within 48 hours, we will send you a reminder.
• Your article will be published Online First approximately one week after receipt of your
corrected proofs. This is the official first publication citable with the DOI. Further changes
are, therefore, not possible.
• The printed version will follow in a forthcoming issue.
Please note
After online publication, subscribers (personal/institutional) to this journal will have access to the
complete article via the DOI using the URL: http://dx.doi.org/[DOI].
If you would like to know when your article has been published online, take advantage of our free
alert service. For registration and further information go to: http://www.springerlink.com.
Due to the electronic nature of the procedure, the manuscript and the original figures will only be
returned to you on special request. When you return your corrections, please inform us if you would
like to have these documents returned.
Metadata of the article that will be visualized in OnlineFirst
ArticleTitle The Collapse of Insensitive Semantics
Article Sub-Title
Article CopyRight Springer Science+Business Media B.V.
(This will be the copyright line in the final PDF)
Journal Name Linguistics and Philosophy
Corresponding Author Family Name Doerge
Particle
Given Name Friedrich Christoph
Suffix
Division
Organization
Address Biesingerstraß 5, Tuebingen, 72070, Germany
Email chris.doerge@hotmail.de
Received
Schedule Revised
Accepted
Abstract The idea motivating their account, Cappelen and Lepore say in Insensitive Semantics (2005), is that semantic
content is context invariant, and that all colleagues who take, or even consider, different accounts are just on
the wrong track. It is the purpose of their book to disprove all alternative accounts by way of an argument
‘by elimination’. The conclusion they arrive at is that their own account must be accepted by everyone as
“the only game in town at the end of the day”. The present paper is intended to examine this conclusion; its
more significant findings are these. Firstly, C&L’s account is not, as they suggest, strictly minimalist, but in
fact just a moderate version of contextualism. Secondly, prematurely associating semantical incompleteness
and context sensitivity, they overlook some possible alternatives to their own view, among them at least one
that is attractive. Thirdly, their argument ‘by elimination’ has an inductive structure, but is inexhaustive, and
therefore inconclusive. Fourthly, for several different reasons, their attempts to reject arguments in favour of
semantical incompleteness do not work. Finally, their contention that arguments in favour of semantical
incompleteness employ metaphysical premisses for semantical conclusions rests on a faulty interpretation of
these arguments. In the light of these findings, it is concluded that the central argument of Insensitive
Semantics fails.
Keywords (separated by '-') Context sensitivity - Insensitive Semantics - Semantic content - Semantical incompleteness - Truth conditions
Footnote Information
Journal: 10988
Article: 9077
Author Query Form
Please ensure you fill out your response to the queries raised below
and return this form along with your corrections
Dear Author
During the process of typesetting your article, the following queries have arisen. Please
check your typeset proof carefully against the queries listed below and mark the
necessary changes either directly on the proof/online grid or in the ‘Author’s response’
area provided below
Query Details required Author’s response

Front Please check and confirm the author
matter names and initials are correct. Also,
kindly confirm the details in the
metadata are correct.
Body Please update the missing text in
matter Acknowledgments and in foot note 29.
Linguist and Philos
DOI 10.1007/s10988-010-9077-y
RESEARCH ARTICLE
The collapse of insensitive semantics

Author Proof
F
Friedrich Christoph Doerge
OO
PR
Ó Springer Science+Business Media B.V. 2010
1 Abstract The idea motivating their account, Cappelen and Lepore say in Insen-
2 sitive Semantics (2005), is that semantic content is context invariant, and that all
3
D
colleagues who take, or even consider, different accounts are just on the wrong
4 track. It is the purpose of their book to disprove all alternative accounts by way of an
5 argument ‘by elimination’. The conclusion they arrive at is that their own account
TE
6 must be accepted by everyone as ‘‘the only game in town at the end of the day’’. The
7 present paper is intended to examine this conclusion; its more significant findings
8 are these. Firstly, C&L’s account is not, as they suggest, strictly minimalist, but in
9 fact just a moderate version of contextualism. Secondly, prematurely associating
EC
10 semantical incompleteness and context sensitivity, they overlook some possible

11 alternatives to their own view, among them at least one that is attractive. Thirdly,
12 their argument ‘by elimination’ has an inductive structure, but is inexhaustive, and
13 therefore inconclusive. Fourthly, for several different reasons, their attempts to
14 reject arguments in favour of semantical incompleteness do not work. Finally, their
RR
15 contention that arguments in favour of semantical incompleteness employ meta-

16 physical premisses for semantical conclusions rests on a faulty interpretation of
17 these arguments. In the light of these findings, it is concluded that the central
18 argument of Insensitive Semantics fails.
CO
19 Keywords Context sensitivity Insensitive Semantics Semantic content

20 Semantical incompleteness Truth conditions
UN
F. C. Doerge (&)
Biesingerstraße 5, 72070 Tuebingen, Germany
e-mail: chris.doerge@hotmail.de
123
Journal : Small-Ext 10988 Dispatch : 8-11-2010 Pages : 24
PIPS No. : 9077 h LE h TYPESET
MS Code : LING422 h CP h DISK
F. C. Doerge
21 1 Introduction
22 ‘‘The idea motivating Semantic Minimalism’’, Cappelen and Lepore (C&L) say in
23 Insensitive Semantics (2005, 143), ‘‘is simple and obvious: The semantic content1 of
Author Proof
24 a sentence S is the content that all utterances of S share’’; it is ‘‘the content that can
25 be grasped and reported by someone who is ignorant about the relevant charac-
F
26 teristics of the context in which an utterance of S took place’’.2 I emphatically agree
27 both to the truth and the significance of this view; following C&L themselves (2005,
OO
28 143), I shall call it the ‘‘Basic Idea’’ of their account.
29 According to the Basic Idea, there is no contribution of context to semantic
30 content at all. What C&L subsequently defend as ‘‘Semantic Minimalism’’, how-
31 ever, is something slightly different: it is the view that there is no contribution of the
32 context to semantic meaning with the exception of members of the set of (genuinely)
PR
33 context sensitive expressions (2005, 17, 151). I shall call this the ‘‘Basic Set’’; its
34 members amount, at least roughly, to those expressions which Kaplan (1989) calls
35 ‘‘demonstratives’’ (cf. C&L 2005, 1f.). Because of the exception they make, calling
36 their view ‘‘Semantic Minimalism’’, as they do themselves, is a bit misleading; in
37 the present paper I shall instead call it ‘‘Very Moderate Contextualism’’ (VMC). A
38 view for which the term ‘minimalist’ is more appropriate will be introduced below.
D
39 But let me first continue presenting C&L’s case.
40 The central task of Insensitive Semantics (IS) is to show that VMC must be true,
TE
41 because all alternative views are unacceptable. In order to achieve it, C&L present
42 what they (2005, 151) call an argument ‘‘by elimination’’. The argument starts from
43 the assumption that there are exactly two alternatives to VMC; one is what they call
44 ‘‘Moderate Contextualism’’, the other is called ‘‘Radical Contextualism’’ (C&L
45 2005, 151, 190). Moderate Contextualism (MC) is the view that, besides demon-
EC
46 stratives, there are other expressions which are semantically incomplete and context
47 sensitive (2005, 7, 39, 59). Radical Contextualism (RC) says that each and every
48 sentence,3 or even each and every expression,4 is context sensitive, and semantically
RR
1
FL01 Cappelen and Lepore (1997, 278) characterise ‘content’ as ‘‘whatever semantic theory attempts to
FL02 elucidate’’. This suggests that on their account ‘semantic content’ refers to a variety of the semantic
FL03 meaning of (declarative) sentences—Kaplan’s (1989) ‘‘character’’, for example, being another variety.
CO
FL04 That their ‘semantic content’ is a kind of semantic meaning is reconfirmed by the ‘Basic Idea’ which is
FL05 just being introduced: for although it is plausible to maintain, about semantic meaning, that it is what
FL06 someone ignorant about the characteristics of the context can know, it is hard to see why such a thing
FL07 should be maintained of ‘semantic content’ if this does not refer to a variety of semantic meaning. An
FL08 alternative interpretation of their account is given, for example, by Korta and Perry (2007, 95f.).
FL09 According to it, ‘semantic content’ is the result of the application of pragmatic impact on semantic
FL10 content, which suggests that ‘semantic content’ is a matter of pragmatics; if we ascribe this notion to C&L
UN
FL11 themselves, however, then their statement of the Basic Idea would be hard to explain.
2
FL01 To say that the context has no significance for the semantic meaning of an expression is, of course, not
FL02 to deny that there may be a contribution of the context to what the speaker means with the expression, or
FL03 what the speaker means in making the utterance, etc.
3
FL01 See, e.g., (2005, 6, n5, 128, 131).
4
FL01 See, e.g., (2005, 6, 14, 31).
123
The Collapse of Insensitive Semantics
49 incomplete.5 So in the main,6 C&L construe the difference between MC and RC as a

50 difference concerning the extension of context sensitivity and semantical incom-
51 pleteness (we turn back to this issue presently).
52 Now as to the argument by elimination, very roughly, it works as follows. Firstly,
Author Proof
53 in C&L’s view RC is a completely unacceptable view. In order to convince us of

54 this, they raise three objections against RC. Firstly, they argue, RC fails the
F
55 so-called ‘‘context sensitivity test’’, thus assuming context sensitivity in cases
56 where, on their own view, there is none.7 Secondly, on their view RC fails to explain
OO
57 the pervasive presence of communicative success; indeed, they argue, if RC was
58 true, then communication would be impossible.8 Thirdly, they argue that RC is an
59 inconsistent view, because assuming it leads to contradictions.9 Thus, they (2005,
60 14, 151) arrive at a devastating assessment of RC: it is ‘‘internally inconsistent’’ and
61 ‘‘empirically flawed’’, and it ‘‘fails to account for some of the basic features of our
PR
62 linguistic practices’’.
63 As to MC, according to C&L it is in no better a position. For although it appears
64 to constitute a more moderate position at first, C&L argue, it finally turns out to
65 force its adherent to the endorsement of RC. C&L construe what they call a
66 ‘‘Slippery Slide Argument’’ (SSA), an argument which sets out to show that all
67 attempts to establish MC entail a commitment to the endorsement of RC, such that
D
68 after all MC cannot be defended without simultaneously endorsing RC. More in
69 particular, the argument proceeds as follows.
TE
70 ‘‘[A]t the center of all arguments that attempt to establish’’ MC, they (2005, 17)
71 argue, are two and only two arguments: ‘‘Context Shifting Arguments’’ and
72 ‘‘Incompleteness Arguments’’.10 Context Shifting Arguments bring to bear intuitions
73 which suggest that ‘‘what is said, or expressed by, or the truth conditions of, an
74 utterance varies in some systematic way with contexts of utterance’’ (2005, 10), thus
EC
75 suggesting that the sentence under consideration is context-sensitive. Incompleteness

76 Arguments, again, employ intuitions to the effect that a certain sentence, or expres-
77 sion, is semantically incomplete (these will be considered in more detail below).
78 What the SSA is designed to show is this: Regardless whether adherents of MC
RR
79 employ an Incompletess Argument, or a Context Shifting Argument, in either case do

80 they ‘‘lead themselves directly into the clutches of RC’’ (2005, 7). Thus, whatever
81 argument is employed in favour of MC, to produce it means to endorse RC.
82 Now given this commitment of arguments in favour of MC to an endorsement of
83 RC, C&L appear to have established their case. MC and RC are the only alternatives
CO
84 to VMC. MC collapses into RC, and thus, they conclude, ‘‘RC [. . .] is the logical
85 consequence of denying’’ VMC (2005, 7). Furthermore, RC appears to be com-
86 pletely unacceptable. Thus, there appears to be no alternative to endorsing VMC. As
UN
5
FL01 They give different characterisations in both respects; see, e.g. (2005, 6, 14, 31, 128, 131; cf. also 6, n5).
6
FL01 They sometimes ascribe additional claims to MC and RC; see, e.g., (2005, 6f).
7
FL01 For details see C&L (2005, Chap. 7).
8
9
10
FL01 They (2005, 10) call this a ‘‘methodological observation’’; see also C&L (2005, 10, 14, 38).
123
F. C. Doerge
87 C&L put it themselves: apparently, VMC remains as ‘‘the only game in town at the
88 end of the day’’ (2005, 151; cf. also 190).
89 In the few years since the publication of their book, a profound number of replies
90 have been given. Indeed, though it sounds appealing at first, most of their opponents
Author Proof
91 remain decidedly reluctant to subscribe to their account and to swallow the con-
92 sequences it brings with it, after all. Still, the central argument of IS, by which they
F
93 set out to show that all other views are unacceptable, has not yet been scrutinised
94 conclusively up to now. The present paper is intended to contribute to this task. It
OO
95 shows that, in several different ways, C&L’s argument fails to work.
96 As a start, in the remainder of this section I shall abandon an assumption which is
97 presupposed by C&L’s argument, but which appears to be unfounded, namely, the
98 strict association of context sensitivity and semantical incompleteness with each
99 other. This view underlies the simple trichotomy which C&L construe, consisting of
PR
100 RC, MC, and VMC. For the claim that these three views exhaust the range of
101 possible accounts makes sense only on the assumption that context sensitivity and
102 semantical incompleteness strictly cooccur. This assumption, however, is highly
103 doubtful.
104 To start with, it is at least not the case that the two terms, ‘‘context sensitivity’’
105 and ‘‘semantical incompleteness’’ are just different words for one and the same
D
106 thing. Nor is this entailed, or even suggested, by the way in which C&L’s describe
107 these terms. According to C&L (2005, 33), context sensitivity is ‘‘variability
TE
108 between contexts’’, and ‘‘to establish such variability one has to think about and
109 compare what happens to the same sentence in different contexts of utterance’’; a
110 sentence is context sensitive when its semantic content ‘‘shift[s] from context to
111 context’’ (2005, 105, cf. 2005, 33, 99, 104).
112 The notion of semantical incompleteness is described quite differently. ‘‘In the
EC
113 case of incompleteness’’, the sentence under consideration ‘‘has [no] conditions of
114 truth, and so, [no] truth value, independent of any context’’ (2005, 33). For without
115 complementation by a context, semantically incomplete sentences express ‘‘only
116 fragmentary propositions’’. Such a sentence lacks truth-conditions because ‘‘it is
RR
117 incomplete qua semantic entity’’ (2005, 7).

118 Thus, when a sentence is semantically incomplete, then this does not necessarily
119 involve any consideration of what might happen to the sentence in one or the other
120 context of utterance; while a sentence’s context sensitivity essentially derives from
121 its behaviour in one or the other particular context. Furthermore, when a sentence is
CO
122 semantically incomplete, then this is derived from the semantical incompleteness of
123 its character, while a sentence’s context sensitivity derives from the context shifting
124 of the semantic contents associated with it in different contexts. Thus, the concept of
125 context sensitive and the concept of semantical incompleteness are different con-
126 cepts, according to the description C&L give of them.
UN
127 It might still be that context sensitivity and semantical incompleteness, though
128 different things, are coextensional. Given that the two ideas are different, however,
129 establishing such a view would seem to require good reasons. Now, on the one hand,
130 C&L themselves do not give any argument to this effect (they do not try to establish
131 this view; they just do not deal with the issue). On the other hand, there are several
132 reasons against coextensionality. Thus, it seems arguable that the expression
123
133 ‘‘Napoleon Buonaparte left in the lurch’’ is semantically incomplete (as well as,
134 additionally, syntactically incomplete), while not being context sensitive. On the
135 other hand, it seems arguable that ‘‘The tallest European is bald’’ is context sen-
136 sitive, but not semantically incomplete.
Author Proof
137 Thus, there is reason not to assume coextensionality of context sensitivity and
138 semantical incompleteness.11 It is then a possible view, and in my opinion indeed an
F
139 attractive one, that some sentences are semantically incomplete, while there is no
140 such thing as context sensitivity—a moderate account of the extension of semantical
OO
141 incompleteness, that means, and a consistently ‘minimal’ account of the extension
142 of context sensitivity.12 This minimal view of context sensitivity, together with a
143 moderate endorsement of semantical incompleteness, can well be combined with
144 different views of the nature and extension of semantic content such as, for example,
145 with the view that not all sentences do have any semantic content in addition to a
PR
146 ‘character’ in Kaplan’s (1989) sense, or with a conception of ‘semantic content’
147 according to which it is possible for a sentence to have a content which is not truth-
148 evaluable, or again, with a rather liberal view of truth-evaluability (similar, for
149 example, to the one C&L defend in IS).13
150 One advantage of the account I am suggesting, at least from the point of view
151 which C&L take, is that it takes their Basic Idea, that semantic content is not context
D
152 sensitive, seriously. For this Basic Idea suggests that the semantic content of a
153 sentence is independent of the context of utterance. By assuming no such thing as
TE
154 context sensitivity, the present view accounts for the Basic Idea straightfor-
155 wardly—while by assuming a considerable amount of context sensitivity,14 C&L’s
156 own theory accounts for their Basic Idea at least less obviously.
157 Furthermore, there is a second advantage of my account from the point of view of
158 C&L’s account. It is a presupposition of their central argument in IS that any
EC
159 account of semantic content which does assume some amount of context sensitivity
160 must defend this view somehow (while, as they argue, any such defence will cause
161 an irresistable collapse into RC, which makes it an unacceptable view after all).
162 Now by suggesting that after all there is some amount of context sensitivity there is
RR
163 reason to believe that their own account eventually succumbs to this problem; while
164 the account I am suggesting, by assuming no context sensitivity at all, perfectly
165 avoids problems of this type—we shall turn back to this issue below.
166 As far as terminology is concerned, since the account I suggest assumes no
167 context sensitivity, I think it is appropriate to call it ‘‘Real Semantic Minimalism’’
CO
168 (Real SM) with respect to context sensitivity. C&L’s own account of context
11
FL01 To be sure, it is not important for the present purpose that non-coextensionality is, or can be proven;
FL02 the crucial thing is that coextensionality cannot be taken for granted.
12
FL01 This minimal view of context sensitivity is similar, e.g., to Carston’s (2008) form of ‘minimalism’ and
UN
FL02 to Korta and Perry’s (2007, 99) ‘KP minimalism’; cf. also Borg’s (2007) conception of ‘minimalism’.
13
FL01 See, e.g., C&L (2005, 2–3).
14
FL01 C&L give different descriptions of how far, on their account, semantical incompleteness and context
FL02 sensitivity are supposed to extend. Initially (2004, 1) they suggest that context sensitivity is restricted to
FL03 the determination of the semantic value of ‘indexicals’ in Kaplan’s sense, but then they weaken the
FL04 restriction by the qualification ‘‘plus or minus a bit’’ (2004, 2); and they later (2004, 144f.) include,
FL05 additionally, disambiguation and precisification.
123
F. C. Doerge
169 sensitivity, however, since it finally does assume some amount of context sensi-
170 tivity, may perhaps more appropriately be identified as ‘‘Very Moderate Contex-
171 tualism’’ (VMC), rather than as ‘‘Semantic Minimalism’’.
172 Anyway, the account I suggest assumes a certain amount of semantical incom-
Author Proof
173 pleteness, while it does not assume any amount of context sensitivity. This ‘mixed’
174 design of the view has certain significant consequences for my examination of the
F
175 argument of IS. As we remember, C&L’s attack against competing accounts
176 involves the rejection of two kinds of phenomena, context sensitivity and semantical
OO
177 incompleteness. Since I suggest taking a view according to which there is no context
178 sensitivity, my account is immune to their arguments against context sensitivity, and
179 thus these arguments need not be taken into consideration here.
180 However, my account does involve a moderate endorsement of semantical
181 incompleteness. C&L call arguments in favour of such an endorsement ‘‘Incom-
PR
182 pleteness Arguments’’. Their relevant SSA says that any Incompleteness Argument,
183 if it is applicable at all, will be applicable to each and every sentence, and that the
184 one who employs such an Incompleteness Argument thus gets committed to the
185 endorsement of RC, which is a completely unacceptable view. After a general
186 analysis of their overall argument in Sect. 2, in Sects. 3 and 4 I shall argue that their
187 SSA against Incompleteness Argument is unsuccessful. It is not true that any
D
188 argument in favour of semantical incompleteness commits us to the view that
189 semantical incompleteness is ubiquitous. As we shall see, there are several reasons
TE
190 for this conclusion. For example, their attempt to reject the two Incompleteness
191 Arguments fails in either case, and each time in more than one way. Additionally,
192 the two Incompleteness Arguments which they take into consideration do not
193 exhaust the range of Incompleteness Arguments, which anyway makes the SSA
194 inconclusive. C&L present another objection to arguments in favour of semantical
EC
195 incompleteness, to the effect that such arguments employ metaphysical assumptions
196 for the support of semantical conclusions, and are therefore unacceptable. In Sect. 5,
197 I examine this line of argument, and show that it is unconvincing, too.
RR
198 2 General analysis of C&L’s argument by elimination
199 So C&L set out to reject all kinds of Incompleteness Arguments: what exactly is an
200 ‘Incompleteness Argument’? Following C&L’s description,15 it is an argument that
CO
201 furthers the view that certain sentences are semantically incomplete. A typical
202 instance of such an argument may, for example, sound like this.
203 Consider the alleged proposition that P that some sentence S semantically
204 expresses. Intuitively, the world can’t just be P simpliciter. The world is
205 neither P nor not P. There’s no such thing as P’s being the case simpliciter.
UN
206 And so, there is no such proposition. (C&L 2005, 11)

207 In general, the goal of an Incompleteness Argument, according to C&L (2005, 11),
208 ‘‘is to establish that the proposition Semantic Minimalists claim is semantically
15
FL01 See C&L (2005, 11, 33–35, 59–61).
123
209 expressed by some sentence S is no more than a propositional fragment’’. That a

210 sentence is ‘semantically incomplete’, means that it must be ‘‘completed’’ by ‘‘add-
211 ing’’ something to its meaning for it to be truth-evaluable, while otherwise this
212 meaning merely constitutes a propositional fragment (2005, 61); and that a sentence
Author Proof
213 expresses a ‘propositional fragment’ means that it ‘‘does not express a complete
214 proposition’’, but instead something that ‘‘cannot have a truth-value’’, and that it
F
215 ‘‘doesn’t express genuine truth conditions’’ (2005, 61).
216 Making their case against Incompleteness Arguments, C&L advance what they
OO
217 call a ‘Slippery Slide Argument’ (SSA). In its most general form, as directed against
218 any version of MC, and following C&L’s (2005, 39, 59) explicit description,16 it can
219 be represented as follows: Given an argument A, if A suffices to show that MC is
220 true, then RC follows. To see exactly what they mean, however, some qualifications
221 are necessary.
PR
222 (1) Taken literally, this statement suggests that the SSAs applies only to arguments
223 which ‘‘suffice to show that’’ MC is true. Construed this way, however, the
224 argument would fail to rebut reasonings which aim at strong evidence for
225 context sensitivity and semantical incompleteness, or at some evidence, rather
226 than at conclusive evidence. Since C&L aim at an exhaustive rejection of all
227
D
arguments in favour of MC, the focus must be wider: the argument is to be
228 construed as directed against any argument providing evidence for MC, as
229 indeed C&L themselves do in other places (e.g., 2005, 14). Thus, the following
TE
230 statement is more accurate: Given an argument A, if A provides evidence in
231 favour of MC, then RC follows.
232 (2) Saying that RC ‘‘follows’’ from arguments in favour of MC, C&L may appear to be
233 claiming an a priori connection between an argument’s supporting MC, and its
EC
234 entailing RC.17 This, however, is not the case. The reason they suggest for the
235 connection is not a priori. Nor do they claim that there is a general reason why
236 arguments in favour of MC commit one to endorsing RC. Their SSA, rather
237 consists of a bunch of partial SSAs, each of them directed individually against one
238 or the other argument in favour of MC, in such a way that for each single argument
RR
239 in favour of MC there is at least one SSA which shows that endorsing this argument
240 commits one to the endorsement of RC. In their own words, they execute the
241 argument by considering particular sets of data ‘‘presented in favour of expanding
242 the Basic Set of context sensitive [sc., and/or semantically incomplete] expres-
243 sions, and show that if this evidence supports an expansion of the Basic Set, then all
CO
244 expressions are context sensitive [sc., and/or semantically incomplete], i.e., RC is
245 true’’ (2005, 14). Thus, the SSA against MC splits up into a more or less huge
246 number of partial SSAs, addressing each and every single argument in favour of
247 context sensitivity and/or semantical incompleteness there is, and arguing in each
248 case that this argument, too, commits one to the endorsement of RC.
UN
249 (3) As a consequence of the inductive structure of the overall SSA, C&L’s case rests
250 upon a variety of particular SSAs, directed against particular arguments in
16
FL01 See also C&L (2005, 14, 60, 61).
17
FL01 This interpretation is also suggested by their view, already mentioned above, that as a consequence of
FL02 SSA, RC is the negation of VMC.
123
F. C. Doerge
251 favour of MC. Since MC has two aspects, context sensitivity and semantical
252 incompleteness, the entirety of arguments for MC can be split in two groups,
253 those concerning context sensitivity, and those concerning semantical incom-
254 pleteness. Since I am defending a moderate view of semantical incompleteness,
Author Proof
255 I shall in the following be concerned with the SSA as directed (or those partial
256 SSAs which are directed) against Incompleteness Arguments.
F
257 Given the fact that C&L give no general reason why arguments in favour of
OO
258 semantical incompleteness are supposed to commit one to the endorsement of RC,
259 and given their contention that each single Incompleteness Argument does involve
260 such a commitment nevertheless, C&L are forced to take each and every single
261 Incompleteness Argument into consideration, and to try to give a reason in each
262 case, why this particular Incompleteness Argument involves a commitment to RC,
263 too.
PR
264 In order for this, it is of course their first task to collect all Incompleteness
265 Arguments there are, in order subsequently to address them one by one. As they put
266 it themselves, ‘‘[f]or us to be able to run [our] kind of argument we need to know
267 [the particular single reasons] why Bach and other Moderate Contextualists think
268 there’s no such thing as these minimal propositions. What makes them incomplete?
269
D
What are the criteria by which one proposition is deemed incomplete and another
270 complete?’’ (C&L 2005, 61).
271 In C&L’s view, the adherent of a moderate contextualism carries the burden of
TE
272 proof: the assumption that a sentence is semantically complete, on their account, is
273 the standard assumption, while the view that a certain sentence is semantically
274 incomplete requires special justification. ‘‘The challenge to the Moderate Contex-
275 tualist’’, they argue, ‘‘is to show how their favorite cases of incompleteness are
EC
276 fundamentally different from any arbitrary English sentence’’ (2005, 65). On the
277 other hand, the inductive structure of their argument also provides a special chal-
278 lenge to their own case: For, evidently, their SSA will work only insofar as they
279 consider really each and every Incompleteness Argument, and only insofar as the
280 SSA they give is really convincing in each case.
RR
281 (4) C&L (2005, 61) ask for ‘‘the criteria by which one proposition is deemed
282 incomplete and another complete’’: one may think that what they thereby have
283 in mind is a definition of ‘semantical incompleteness’.18 This, however, is
284 probably not what they mean. For in fact they do have and apply a sufficiently
CO
285 clear definition of ‘semantical (in)completeness’. That a sentence is incomplete,

286 on their account, means that it is ‘‘semantically underdeterminate’’ (2005, 34) in
287 such a way that it ‘‘is not fully propositional [. . .]; it is not truth-evaluable’’
288 (2005, 7); ‘‘it is not the kind of thing that can take a truth value’’ (2005, 59).19
UN
18
FL01 Cf. also (2005, 62 n1), where C&L deplore the lack of ‘‘any sharp definition of ‘in/complete’ in [the]
FL02 literature’’.
19
FL01 For sufficiently clear conceptions of the notion of semantical incompleteness see also, e.g., Bach
FL02 (1994, 126f.), or Sperber and Wilson (1986, 188).
123
289 What C&L probably have in mind, when they demand ‘‘criteria’’ of (in)com-
290 pleteness, is some kind of evidence for, or an indication of the semantical
291 incompleteness of sentences.
292 (5) As we saw, because of the inductive structure of the (overall) SSA, its success
Author Proof
293 depends on an exhaustive rejection of all Incompleteness Arguments (as well as,
294 additionally, of all arguments in favour of context sensitivity). Now in their
F
295 discussion, C&L take exactly two Incompleteness Arguments into consider-
296 ation, to which we shall turn presently. Since they end up claiming that their
OO
297 case is successful, they are committed to the assumption that these two argu-
298 ments exhaust the range of Incompleteness Arguments. For if there were
299 additional Incompleteness Arguments, then the inductive procedure would be
300 incomplete, and the conclusion, that the case is successful, underdetermined.
301 Now let us turn to an analysis of those two Incompleteness Arguments. The first of
PR
302 them may be called the ‘‘Open Question test’’ argument. It takes the fact that a given
303 sentence, as it stands, leaves certain questions unanswered as an indication that this
304 sentence is semantically incomplete. Schematically, it can be stated as follows.
305 The ‘‘Open Question test’’ argument:
307 (1) Some sentences are subject to Open Questions.
D
308 (2) Open Questions indicate semantical incompleteness.
309 Hence,
TE
310 (3) There is reason to believe that some sentences are semantically incomplete.
311 The second Incompleteness Argument refers to a criterion which C&L ascribe to
EC
312 Ken Taylor, and which produces the ‘‘felt inability to evaluate the truth value’’ of a
313 sentence (the ‘‘Felt Inability’’, for short) as an indication of this sentence’s
314 semantical incompleteness. Schematically, it can be stated as follows.
315 The ‘‘Felt Inability test’’ argument:
RR
317 (4) Some sentences trigger a Felt Inability.

318 (5) Felt Inabilities indicate semantical incompleteness.
319 Hence,
CO
321 C&L’s argument entails that each and every Incompleteness Argument commits
322 one to endorsing RC, and they assume that the range of Incompleteness Arguments
323 consists of these two arguments. Accordingly, they set out to construe for either of
324 them an SSA, intended to show that employing this Incompleteness Argument
UN
325 commits one to the endorsement of RC. Let us turn to an examination of the two
326 SSAs, starting with the SSA against the Open Question test argument.
19
FL03 For sufficiently clear conceptions of the notion of semantical incompleteness see also, e.g., Bach
FL04 (1994, 126f.), or Sperber and Wilson (1986, 188).
123
F. C. Doerge
327 3 Analysis and defence of the ‘Open Question test’ argument
328 C&L start considering the Open Question test with reference to three groups of
329 examples, originating from authors who, as they (2005, 62) suppose, are adherents
Author Proof
330 of MC, such as Bach (1994b, 128ff.), Carston (1988, 167), Sperber and Wilson
331 (1986, 188), Récanati (1993, 235), Perry (1993, 206) and Crimmins (1992, 17). The
F
332 argument they develop with reference to these examples proceeds in two stages.
333 The first stage refers to an example which Kent Bach uses in order to illustrate how
OO
334 the Open Question test works:
336 (6) ‘‘Steel isn’t strong enough’’
337 This sentence, according to Bach (1994b, 127), is ‘‘semantically or conceptually

338 incomplete’’. ‘‘The conventional meaning of the sentence’’, he argues, ‘‘expresses
PR
339 not a full proposition but merely a propositional radical’’, which entails that the
340 sentence, as it stands, is not truth-evaluable. Rather, ‘‘for the sentence to express a
341 complete and determinate proposition (something capable of being true or false)’’,
342 something would have to be ‘‘added’’. ‘‘[A] complete proposition would be
343 expressed, a truth condition determined’’, by the sentence, ‘‘only if [it] were elab-
344
D
orated somehow’’ (ibid.).
345 Now if we assume that we must, for one or the other reason, end up with a
346 ‘semantically complete’, truth-evaluable proposition, we may say that ‘‘the utter-
TE
347 ance of a semantically underdeterminate sentence requires completion’’, or that a
348 certain constituent ‘‘is needed to complete the proposition’’ (ibid.). And to indicate
349 just what this constituent is that ‘‘must be added for the sentence to express a
350 complete and determinate proposition’’, we can refer to the fact that a certain
EC
351 question ‘must be asked’, and answered, in order to arrive at a semantically com-
352 plete, truth-evaluable proposition. Thus we can say, for instance, that the sentence
353 ‘‘Steel isn’t strong enough’’ is semantically incomplete, and specify the way in
354 which it is, by saying that ‘‘we need to know strong enough for what’’ (ibid.).
355 In their objection against Bach’s argument, C&L introduce a counterpart to
RR
356 Bach’s example, enriched by the words ‘‘to support the roof’’, which are supposed
357 to bridge the gap diagnosed by the Open Question:
359 (7) ‘‘Steel isn’t strong enough to support the roof’’
CO
360 Although this complementation is intended to bridge the gap indicated by the
361 Open Question, C&L (2005, 63) argue, it does not achieve this task, as far as
362 semantical incompleteness is concerned. For in fact, (7) is ‘‘equally susceptible to
363 Incompleteness Arguments’’ as (6) is (2005, 62). It does not, for example, ‘‘settle for
364 how long the support must last. Do a few seconds suffice? More than 3 days? Many
UN
365 days? Many years?’’ (2005, 63). Indeed, they (2005, 62) argue, there is no ‘‘prin-
366 cipled difference’’ between (6) and (7). For ‘‘one can trigger exactly the same kind
367 of incompleteness intuitions’’ for (7) as those advanced for (6) (ibid.).
368 Now as far as this is concerned, it is not quite certain that the adherent of MC will
369 be particularly impressed. At least it is easy to see several ways in which she may
370 react to this argument without conceding ground. For example, she may complain
123
371 that the complement which C&L offer, ‘‘to support the roof’’, is not a very prom-
372 ising device for the task of bridging the gap indicated by the question ‘‘Strong
373 enough for what?’’: if ‘‘to support the roof’’ raises additional questions such as ‘‘For
374 how long’’, then this only shows that it is not suitable for the task it is intended to
Author Proof
375 achieve. Also, the adherent of MC may probably insist that even if such a question
376 as ‘‘For how long?’’ remains unanswered, it is easy to provide another sentence, in
F
377 which the remaining gap is bridged.
378 But now the second stage of C&L’s argument comes, and this threatens to
OO
379 steamroll any attempt to defend MC along these lines. It proposes that each and
380 every sentence whatsoever is subject to one or more incompleteness-indicating
381 questions. The idea underlying this second stage is simply to adopt an argument by
382 Charles Travis, whom C&L see as ‘‘an ally in this context’’, because ‘‘he, as a
383 Radical Contextualist, is also in the business of showing that there’s nothing special
PR
384 about’’ such a sentence as ‘‘Steel isn’t strong enough’’ (2005, 62).
385 Travis’ argument is intended to show that the assessment of sentences as true or
386 false meets more obstacles than one would at first expect. Take, for instance, the
387 sentence ‘‘The kettle is black’’. It may be uttered in different circumstances,
388 including these:D
389 Suppose that the kettle is normal aluminium, but soot covered; normal alu-
390 minium but painted; cast iron, but glowing from heat; cast iron but enamelled
391 white on the inside; on the outside; cast iron with a lot of brown grease stains
TE
392 on the outside; etc. (Travis 1985, 197)
393 If the sentence is uttered in circumstances like these, then it appears to be
394 impossible to say what exactly it is that ‘is said’, or at least impossible to say whether
395 ‘what is said’ is true or false (Travis 1985, 196f). As a consequence, if such a
EC
396 sentence is uttered in such circumstances, then either ‘‘it remains indeterminate
397 precisely what was said in [these circumstances], though perhaps what was said is
398 determinably true or false, if we only knew what it was’’, or else ‘‘something is said
399 which is definitely truth valueless under the conditions depicted’’ (Travis 1985, 198).
RR
400 The kind of questions which Travis introduces here (in the following we call
401 them Travis-questions, and the test referring to them the Travis-question test) is
402 similar to, although not quite identical with, the Open Question test which Bach
403 applies. What is examined by a Travis-question is the semantic content of a sentence
404 in a given context, and the Travis-question test confronts a given sentence with the
CO
405 question of whether it is true or false in the face of one or the other special context
406 of utterance.
407 Travis suggests that either there is no determinate semantic content in this
408 context, or the semantic content of the sentence in such a context is such that the
409 sentence cannot be assessed as either true or false. Furthermore, Travis suggests that
UN
410 for many sentences such special contexts can be construed, contexts in which either
411 there is no determinate semantic content, or in which the semantic content cannot be
412 assessed as either true or false.
413 C&L apply Travis’s argument in order to suggest that the application of Open
414 Question tests commits one to a radical view of semantical incompleteness. The
415 crucial fact (or supposed fact) they refer to is that the unability to say whether the
123
F. C. Doerge
416 sentence is true or false in one or the other special context can be elicited, not
417 merely with some sentences, or many sentences, but indeed with each and every
418 possible sentence. On C&L’s view, those adherents of a moderate view of incom-
419 pleteness who have the intuition that some sentences are truth-evaluable as they
Author Proof
420 stand, while others, such as ‘‘She’s ready’’ or ‘‘It’s raining’’, are semantically
421 incomplete, overlook how universally the Travis-test turns out to be applicable,
F
422 given only some imaginative power. ‘‘[I]f the reason why [sentences] are incom-
423 plete is because they fail to answer these kinds of ‘would it be true if . . .’-ques-
OO
424 tions’’’, C&L (2005, 63) ask, then ‘‘why should we hold that this alleged
425 incompleteness is limited to some (more or less circumscribed subset of [. . .]20
426 sentences?’’
427 For although, as they admit, ‘‘[y]ou might initially feel that [a sentence like
428 ‘‘Smith weighs 80 kg’’] requires contextual supplementation, not in the same way as
PR
429 ‘She’s ready’ and ‘It’s raining’ do’’, ‘‘here’s a prediction: If you read a few of
430 Travis’ little stories, you will eventually have that feeling’’ (2005, 67). In fact, they
431 contend, each and every sentence can be confronted with Travis-questions. ‘‘What
432 Radical Contextualists see clearly, and are prepared to bravely pursue to the bitter
433 end’’, they argue, ‘‘is that the same sort of incompleteness is to be found in every
434 English sentence.’’ (2005, 63). From the applicability of Travis-questions to all
D
435 sentences (rather than merely to some, or many), C&L conclude that if the Open
436 Question test argument is accepted at all, then it will suggest that all sentences
TE
437 (rather than just some, or many) are semantically incomplete. Thus they (2005, 68)
438 arrive at the view that if an argument like the Open Question test argument is
439 applied at all, then it ‘‘also follows that RC is true’’.
440 Some of the statements with which C&L advance Travis’ argument may sound as
441 if they actually believed that all sentences are semantically incomplete. However,
EC
442 they are at great pains not to subscribe to this view.21 Instead, their argument
443 continues with an attempted reductio ad absurdum. As C&L claim to show in their
444 book, RC is quite an unacceptable view. Hence, although the Travis-test, on the
445 assumption that it is significant, provides plausible evidence for the incompleteness
RR
446 of all sentences, this must not be taken as evidence that all sentences are seman-
447 tically incomplete, but rather as evidence that the Open Question test must be
448 treacherous.
449 Let us summarise the crucial elements of the argument schematically. The
450 argument examines the validity of assumption (2), on which the Open Question test
CO
451 argument is based:

453 (2) Open Questions indicate semantical incompleteness.
454 This assumption is challenged in the following way:

UN
456 (8) All sentences are subject to Travis Questions.
20
FL01 ‘‘English’’, they qualify here; however, being about meaning in general their argument should not be
FL02 restricted to English sentences and meanings.
21
FL01 Cf., e.g., C&L (2005, 59).
123
457 (9) Travis-questions are a variety of Open Questions.

458 Hence, . . .
459 (10) On the assumption that Open Questions indicate semantical incomplete
Author Proof
ness, (2) implies RC/commits one to RC.
F
460 The conclusion represented in (10) is supposed to be devastating for a moderate
461 view of semantical incompleteness, for the reason that the argument can be con-
OO
462 tinued as follows:
464 ...
465 (11) RC is an absurd view.
466 Hence, . . .
PR
467 (12) The assumption that Open Questions indicate semantical incompleteness
implies an absurd view.
468 Before we turn to an assessment of the SSA, it should be noticed that C&L’s
469 analysis of the Open Question test argument appears to be infected by an overly
470 liberal interpretation of the Open Question test itself. Thus, for example, some of the
D
471 ‘Open Questions’ they construe for the sentence ‘‘John went to the gym’’ are not of
472 the appropriate kind. Such questions as ‘‘Did something in the gym?’’, ‘‘Did what in
TE
473 the gym?’’, and ‘‘For how long?’’, which C&L introduce, do not indicate any
474 intuition of incompleteness of the sentence ‘‘John went to the gym’’: regardless of
475 whether the sentence is semantically complete or not, these questions do not mark
476 any special variety of (felt) incompleteness.
EC
477 To see what went wrong, it is helpful to see exactly what C&L (2005, 64) say
478 about these questions: they argue that they can be asked—rather than that they must
479 be asked and answered. That they can be asked, however, does not make them
480 ‘Open Questions’ in the sense under consideration, as Kent Bach,22 rightly
481 emphasises; for these are question which have to be answered, in order for the
RR
482 sentence to be truth-evaluable. Accordingly, it is quite pointless, as far as the truth-

483 conditions of ‘‘John went to the gym’’ are concerned, to ask what John did in the
484 gym. Whether he did go or not does not depend on how this question is answered.
485 Now while those questions which can be stated are probably irrelevant, those which
486 need to be answered are relevant for our issue. In the following I shall call questions
CO
487 of the relevant variety ‘‘incompleteness indicating questions’’ (IIQs).

488 Now let me present my objections to the SSA against the Open Question test.
489 Two of my arguments are reasons not to believe premise (8), that each and every
490 sentence is subject to one or the other Travis-question, and the third is directed
491 against the truth of premise (9), that Travis-questions are a variety of those Open
UN
492 Questions employed by some defenders of MC.

493 The main reason C&L provide for endorsing this assumption is confidence in the
494 pervasive applicability of Travis-questions. Against the one who assumes that there
495 are semantically complete sentences, they (2005, 67) direct the contention that they
22
FL01 Id., 2006, 441.
123
F. C. Doerge
496 are able to trigger incompleteness intuitions with respect to each and every sen-
497 tence—‘‘at least if we are allowed to talk to people about the relevant feelings’’; and
498 to the one who is still not persuaded, they make ‘‘a prediction: If you read a few of
499 Travis’ little stories, you will eventually have’’ the feeling of incompleteness with
Author Proof
500 every sentence whatsoever (2005, 67).

501 However, considering the matter more thoroughly, some reasons emerge for the
F
502 view that, even if they are allowed to talk to us,23 we should better remain skeptical.
503 As a first reason, I should like to highlight the implausibility of what might be called
OO
504 an endless chain of Travis-questions and complements. If a sentence is confronted
505 with a Travis-question, then in general there seems to be no obstacle to construing a
506 sentence which is complemented in such a way as to avoid the question. Thus, if
507 ‘‘Smith weighs 80 kg’’ is confronted with the question ‘‘Do we undress Smith?’’,
508 then we can provide ‘‘Smith weighs 80 kg undressed’’ as a new candidate of a
PR
509 semantically complete sentence.
510 C&L will not hesitate to insist that for this sentence, again, some IIQ can be
511 construed; let us take their own example: ‘‘Smith weighing 80 kg when?’’. But again
512 a sentence can be construed which resembles ‘‘Smith weighs 80 kg undressed’’, but
513 is complemented in such a way as to avoid the question, such as ‘‘Smith weighs 80
514 kg at 10:00 am MET, undressed’’. To be sure, I have no reason to claim that this
D
515 sentence is semantically complete after all. The point is merely that even if there is
516 another Travis-question for it, again a new sentence can be construed which fills the
TE
517 gap indicated by it. Now when C&L are claiming that each and every sentence is
518 subject to some (apparently) IIQ, then they commit themselves to the position that
519 this procedure, of inventing a new Travis-question, and of then subsequently con-
520 struing a less incomplete sentence, can never end. But this idea does not sound very
521 plausible after all, does it. If it is false, however, then there is no endless chain, and
EC
522 if there is no endless chain, then applying Open Question tests does not commit one
523 to the endorsement of RC: the SSA against the Open Question test fails.
524 There is a second, and even straighter reason why the SSA against the Open
525 Question test does not work. Consider plaitudes and absurdities, necessary truths
RR
526 and necessary falsities. If, for example, a sentence is necessarily false, then it is
527 false in all contexts; thus, there can be no context in which it is not false, and hence
528 no context in which its truth or falsity cannot be evaluated. Consider ‘‘The first
529 human on the moon was a car manufactured by the ancient Greeks’’: although a
530 speaker, in uttering the sentence, may mean one or the other thing which is true,
CO
531 false, or neither of both in itself, the sentence is false, whatever the context of
532 utterance. Therefore, whatever the context, there is no IIQ for it. Similar examples
533 are ‘‘Napoleon I., emperor of France, was a green prime number’’, or ‘‘Contrary to
534 the facts, there is no difference between you and your mother’’. Or consider
535 ‘‘2 þ 2 ¼ 7’’: since it is necessarily false, it must be false under all conditions.
UN
536 Therefore, however fancyful a context is construed by either Travis or C&L, we can
23
FL01 I did ask them to talk to me: I gave them the examples used in the following, among others, and
FL02 requested them to give arguments why they think they are semantically incomplete, giving an Open
FL03 Question for them, and eliciting in me a Felt Inability concerning them. Ernie Lepore said it was too long
FL04 ago, and I should better ask Cappelen. Cappelen did not deal with the actual task at all, even after
FL05 I repeated my request.
123
537 be sure that the sentence is false in it; thus, we can be sure that there will be no
538 Travis-question for it. Similar examples are ‘‘Some bachelors are unmarried’’, or
539 ‘‘Some prime numbers are even numbers’’.24
540 The third reason why the SSA against the Open Question test does not work is
Author Proof
541 internal: the argument rests upon a confusion. C&L try to convince us that each and
542 every sentence is subject to an Open Question by arguing that all sentences are subject
F
543 to a Travis-question. For the success of this argument it is essential that assumption
544 (9), that Travis-questions are a variety of Open Questions, holds true. However, (9) is
OO
545 false. To start with, the Open Question test examines a sentence as it stands, without
546 consideration of any particular context; while the Travis-question test concerns a
547 sentence as confronted with a certain special context. Furthermore, if a sentence
548 provokes an Open Question, then the incompleteness indicated is an incompleteness
549 of its character; while in the case of a Travis-question test, the incompleteness
PR
550 indicated, depending on the presence of the special features of the context mentioned
551 by the question, concerns the semantic content in such a context. Additionally, the
552 semantical incompleteness indicated by an Open Question is categorical, while the
553 semantical incompleteness indicated by a Travis-question is not. In the case of a
554 Travis-question you cannot say ‘‘The sentence lacks truth conditions’’, full stop.
555 When the kettle is made of aluminium and soot covered, then it may be hard, or
D
556 impossible, to say whether ‘‘The kettle is black’’ is true. Consider, however, the case
557 where the kettle is made of plastics and black through and through; in such a case, the
TE
558 availability of the Travis-question does not prevent the sentence from being true. Or
559 consider the case where the kettle is green: in such a case, the sentence is unam-
560 biguously false. The presence of the Travis-question does not prevent the sentence
561 from having conditions of truth and conditions of falsity; it merely shows that in
562 addition there are conditions where the sentence’s truth-value is indeterminable.25
EC
563 Arguing that the pervasive availability of Travis-questions entails the universal
564 applicability of Open Questions, C&L apparently overlook the difference between
565 Travis-questions and those Open Questions which, for example, Kent Bach is
566 concerned with. Even if, as C&L contend, each and every sentence provokes one or
RR
567 the other Travis-question (which, as we saw, is probably not true anyway), this does
568 not entail that the Open Question test argument commits one to the endorsement of
569 ubiquitous semantical incompleteness, because Travis-questions are not a variety of
570 Open Questions.
CO
571 4 Analysis and defence of the ‘Felt Inability test’ argument
572 Let us turn to the second indicator of semantical incompleteness, the ‘‘Felt
573 Inability’’. Taylor introduces the idea with reference to the sentence ‘‘It’s raining’’.
UN
24
FL01 Again, what a speaker means in uttering those sentences may be true or false or neither: but what the
FL02 speaker means is to be distinguished from what the sentence means.
25
FL01 The situation in the case of Travis-questions, as truth conditions are concerned, appears to be quite
FL02 much the same as with the traditional analysis of ‘‘The present king of France is bald’’: if there is no king
FL03 of France, then we have failure of presupposition and some such problem as lack of truth value; while if
FL04 there is a king of France, then this problem disappears.
123
F. C. Doerge
574 ‘‘[I]n the absence of a contextually provided location’’, he (2001, 53) argues, there is
575 ‘‘a felt inability to evaluate the truth of an utterance’’ of this sentence. He suggests
576 taking this as an indication that the sentence is semantically incomplete. Above, we
577 stated the argument as follows:
Author Proof
578 The ‘‘Felt Inability test’’ argument:
F
580 (4) Some sentences trigger a Felt Inability.
OO
581 (5) Felt Inabilities indicate semantical incompleteness.
582 Hence,
584 Before turning to the SSA against this argument, let us deal with an additional
PR
585 complaint C&L (2005, 67) produce against it: there is no reason to assume, they
586 argue, that ‘‘psychological facts about how people feel have any bearing whatsoever
587 on the metaphysical question of whether a proposition exists (or is complete)’’. This
588 objection fails, and obviously so, because it misrepresents the ‘‘Felt Inability test’’ in
589 (at least) two ways. Firstly, the conclusion aimed at in Taylor’s argument is not, as
590
D
they tacitly assume, a metaphysical statement, but a semantical statement. We must
591 distinguish the claim that, given a certain possible thing (proposition), this possible
592 thing (proposition) exists, or does not exist, from the claim that the meaning of a
TE
593 given sentence is, or is not, truth-evaluable. The first claim may be a matter of
594 metaphysics, but it is not the one under consideration here. What is at issue is the
595 second: but this is quite an innocent semantical claim (we turn back to this issue
596 below).
EC
597 Secondly, regardless of whether the Felt Inability is, or involves, a ‘‘psycho-
598 logical fact’’ or not, it is at least additionally a statement about linguistic intuition.
599 Someone’s feeling unable to evaluate the truth of a given sentence in the absence of
600 additional information will standardly involve, among other things, her having one
601 or the other conception of what the sentence means. If she is a (more or less)
RR
602 competent speaker of English, then it is not unreasonable to take her intuition that
603 the sentence is not truth-evaluable as (more or less strong) evidence that it is not,
604 and to take her feeling that additional information would help as evidence that we
605 are confronted with semantical incompleteness. Taking this into consideration, we
606 see that C&L’s complaint is insignificant: for after all, the Felt Inabilty test argu-
CO
607 ment does not, as they assume, refer to psychological facts to support metaphysical
608 conclusions; it rather employs reference to linguistic intuitions to support linguistic
609 conclusions.
610 Let us turn to C&L’s chief objection to the Felt Inability test, the SSA.26 ‘‘[I]t is
611 possible’’, they (2005, 67) argue, ‘‘to trigger [a Felt Inability] with respect to any
UN
612 sentence whatsoever’’. In order to convince us, they employ two examples, con-
613 fronting each of them with several Travis-questions. The first example is ‘‘Smith
614 weighs 80 kg’’. Although the meaning of this sentence may at first appear to be
26
FL01 As with the Open Question test argument, the argument goes back to thoughts by Charles Travis (see,
FL02 e.g., id. 1996); Travis himself, however, does not state the argument as explicitly as C&L do.
123
615 truth-evaluable, they argue, this appearance can be unsettled by means of questions
616 such as ‘‘Smith weighing 80 kg when?’’, ‘‘Under what conditions are the weighings
617 done?’’, ‘‘Do we undress Smith?’’ (2005, 67). The same holds for the sentence
618 ‘‘That’s a red apple’’, considering questions such as ‘‘Red where?’’, ‘‘In what
Author Proof
619 light?’’, ‘‘Scrubbed?’’, ‘‘Brushed?’’, ‘‘Painted?’’ (2005, 68). In the face of cases like
620 these, C&L conclude that ‘‘there’s no reason to think [that Felt Inabilities] are
F
621 triggered with respect to just a limited subset of English sentences’’ (2005, 68).
622 Rather, ‘‘it is possible to trigger these feelings with respect to any sentence what-
OO
623 soever, at least if we are allowed to talk to people about the relevant feelings’’, they
624 (2005, 67) contend.
625 As in the case of the Open Question test, the clue of the SSA is the (allegedly)
626 ubiquitous applicability of Travis-questions; in the present case this is connected
627 with the assumption that Travis-questions trigger a Felt Inabilty; the assumption that
PR
628 Felt Inabilities indicate semantical incompleteness thus appears to lead to RC. More
629 precisely, the structure of the argument is this.
631 (13) On assumption (5) of the Felt Inability test argument, that a Felt Inability
indicates the semantical incompleteness of the sentence considered, a
moderate view concerning the extension of semantical incompleteness is
D
indicated if some but not all sentences trigger a Felt Inability, while a radical
view is indicated if all sentences do so.
TE
632 (14) Each and every sentence allows for Travis-questions, and Travis-questions
always trigger a Felt Inability.
633 Hence, . . .
EC
634 (15) On assumption (5) of the Felt Inability test argument, a radical view of the
extension of semantical incompleteness is indicated.
635 (The argument proceeds as in the case of the Open Question test argument.)
RR
636 The reasons why this SSA does not work resemble those introduced in the case of
637 the Open Question test argument—which is not surprising if we take into consid-
638 eration that again the clue of the argument is the allegedly ubiquitous applicability
639 of Travis-questions. Thus, the first reason why the SSA fails is the implausibility of
640 an endless chain of Travis-questions and complements. Assuming that the defender
CO
641 of MC reacts to each Travis-question by construing a sentence which is less

642 incomplete, and the defender of VMC continues each time trying to invent a new
643 Travis-question, it is not very plausible to assume that the chain will continue
644 strictly ad infinitum. But if it does not, then this suggests that premise (14) is not
645 true, which leads to the breakdown of the argument.
UN
646 The second way in which the SSA fails was already mentioned with the Open
647 Question test argument, too. There are sentences which appear to be truth-evaluable
648 in each and every context, just because their truth or falsity holds trivially, or
649 necessarily. This is good reason to believe that C&L, even if they are allowed to talk
650 to us, will be unable to construe any Travis-question for them. But if there really are
123
F. C. Doerge
651 such sentences, which are not subject to any Travis-questions, then there is a second
652 reason why the SSA against the Felt Inability test argument does not work.
653 There is a third reason. The SSA rests on the assumption that Travis-tests trigger
654 a Felt Inability, which is used to suggest that using the Felt Inability as an indication
Author Proof
655 of semantical incompleteness leads to the conclusion that all sentences are
656 semantically incomplete. However, this assumption is false. Although it may be that
F
657 Travis-tests trigger a felt inability to evaluate the truth of a sentence, what they
658 trigger is not a Felt Inability of the kind under consideration here. (For obvious
OO
659 reasons, the argument resembles the third argument in the case of the Open
660 Question test argument.)
661 Firstly, the Felt Inability which the defender of MC is supposed to be concerned
662 with is triggered by considering a sentence as it stands, without considering one or
663 the other special context. The feeling triggered by Travis-questions, in contrast, is
PR
664 triggered by a sentence as applied to one or the other special context. Secondly, the
665 feeling triggered by a Travis-question indicates semantical incompleteness of the
666 semantic content of the sentence in a particular context; the Felt Inability referred to
667 by the defender of MC, however, is supposed to indicate the semantical incom-
668 pleteness of the character of the sentence under consideration, without consider-
669 ation of any particular context. Thirdly, the Felt Inability which the defender of MC
D
670 is concerned with indicates categorical incompleteness; the Travis-test, however,
671 does not; for a sentence may lack truth-conditions under the special circumstances
TE
672 brought in by the Travis-question, while still having truth conditions in other
673 contexts. Thus, taking C&L’s example, the felt inability to evaluate the truth of
674 ‘‘That’s a red apple’’, as confronted with the question ‘‘Red where?’’, does not
675 indicate that the sentence has no truth conditions at all, full stop; it merely indicates
676 that it lacks them in case the apple is red somewhere, but not everywhere. The felt
EC
677 inability triggered by the question does, for instance, not tell against the view that if
678 the apple is red through and through, then the sentence is true, or that if the apple is
679 green with brown spots, the sentence is false.
680 Thus, after all the SSA against the Felt Inability test argument fails for at least
RR
681 two reasons. Firstly, the assumption it makes, that every sentence is subject to
682 Travis-questions, is false. Secondly, the felt inability triggered by Travis-questions
683 is not of the type under consideration in the Felt Inability test argument, such that
684 even if every sentence were subject to Travis-questions, there would be no danger
685 that RC follows from the application of the Felt Inability test argument.
CO
686 So the SSA against Incompleteness Arguments fails, because the SSAs against
687 both Incompleteness Arguments do not work, for more than one reason in either
688 case. Over and above this, however, there is another reason why the SSA against
689 Incompleteness Arguments fails. As was emphasised above, the (overall) SSA
690 depends in its soundness on the assumption that the two Incompleteness Arguments
UN
691 which C&L take into consideration, the Open Question test argument and the Felt
692 Inabilty test argument, really exhaust the range of Incompleteness Arguments.
693 However, this is not the case: additional Incompleteness Arguments—additional
694 arguments in favour of semantical incompleteness—can be construed.
695 We may, for example, take the impossibility to evaluate a sentence’s truth (rather
696 than the feeling that evaluating it is impossible, as in the case of the Felt Inability
123
697 test) to be an indication of its semantical incompleteness, given that, in addition, the
698 sentence does not support the derivation of any trivial falsities. This test might be
699 built on the assumption that there are exactly two occasions for non-truth-evalu-
700 ability, semantical incompleteness and self-contradictoriness. To be sure, for my
Author Proof
701 argument it is not important that this test is quite appropriate; it would fully suffice
702 that it, or something like it, can be used to build another Incompleteness Argument.
F
703 Another possible indicator of semantical incompleteness (deriving from an
704 incompleteness of character), and one which C&L actually may be expected to
OO
705 appreciate indeed, is a sentence’s containing one or more indexical expressions.
706 Thus one may assume that while ‘‘She’s happy’’ is semantically incomplete,
707 ‘‘Napoleon I. Buonaparte is happy’’ is semantically complete, or at least semanti-
708 cally incomplete to a lesser extent. Yet a third possible indicator of semantical
709 incompleteness neglected by C&L is the presence of vague expressions (in the sense
PR
710 mentioned by C&L 2005, 145). Thus, for whatever deeper reason, one may suppose
711 that compared with ‘‘Peter is huge’’, ‘‘Peter is 1,80 m tall’’ is either not semantically
712 incomplete at all, or at least less incomplete.
713 These three hints are, of course, not meant to give any systematic account of
714 Incompleteness Arguments, or indicators of semantical incompleteness. They do
715 suffice, however, to show that the Open Question test argument and the Felt
D
716 Inability test argument cannot be passed off as the only Incompleteness Arguments.
717 The range of possible arguments in favour of a moderate view concerning
TE
718 semantical incompleteness is less poor than C&L believe. Given this, and given the
719 inductive structure of the SSA, however, there is another way in which the SSA
720 against incompleteness Arguments turns out to be inconclusive.
721 What is more, pursuing the issue further on, things turn out to be even worse for
722 C&L’s argument. Consider what happens when C&L try to strengthen their
EC
723 inductive reasoning by really taking all Incompleteness Arguments into account. Let
724 us, as an example, consider the second case just mentioned, where the presence of
725 indexical expressions is taken as an indication of semantical incompleteness. In
726 order to avoid the breakdown of their SSA, C&L have to construe an argument to
RR
727 the effect that if we take indexical expressions as an indication of semantical

728 incompleteness, then RC concerning semantical incompleteness follows. Further-
729 more, there is a Context Shifting Argument to the effect that sentences containing
730 indexicals are context sensitive, too, and thus C&L will additionally be forced to
731 construe and subscribe to an argument to the effect that such a view leads to
CO
732 radicalism concerning context sensitivity. However, in fact they cannot consistently
733 subscribe to arguments like these. For in fact their own theory assumes the presence
734 of semantical incompleteness and context sensitivity in cases involving indexicals.
735 Following the rationale of their objection, this would commit them to giving some
736 argument in favour of context sensitivity and semantical incompleteness in the case
UN
737 of indexical expressions. Now if such arguments did commit one to RC, then C&L
738 would, following their own SSA, be comitted to the endorsement of RC, and insofar
739 it is true that RC is quite unacceptable, as they continue arguing, they would commit
740 themselves to quite an unacceptable view. In short, the attempt to save their SSA by
741 completing the inductive argument would force them to disprove their own theory
742 by this very SSA.
123
F. C. Doerge
743 The general problem underlying this dilemma has to do with the structure of
744 C&L’s own account, which suffers from an ambiguity between contextualism and
745 non-contextualism. While their SSA against MC pretends that their views are
746 decidedly anti-contextualist, and although they call their view ‘‘semantic mini-
Author Proof
747 malism’’ and present it as an (or rather, the) alternative to ‘‘contextualist’’ accounts,
748 their own theory does contain several contextualist elements, too. Thus, as we saw,
F
749 they do assume context sensitivity and/or semantical incompleteness in the case of
750 indexical expressions, vague expressions, and imprecise expressions. Given this
OO
751 ambiguous structure of their account, the SSA cannot be showing, as it is intended
752 to, that all contextualist accounts fails while C&L’s account lives: any argument
753 which really provides an exhaustive rejection of all contextualist elements will
754 additionally be disproving C&L’s own account, too.
PR
755 5 Are Incompleteness Arguments metaphysical arguments?
756 Before turning to C&L’s third main argument, let us briefly consider a remark
757 which they make quite much in passing, and which may appear as critically directed
758 against Incompleteness Arguments, too, but which is in no way worked out, and
D
759 which in any case can be dismissed rather easily. Both the Open Question test and
760 the Felt Inability test appear to involve intuition, though in slightly different ways.
TE
761 Now in the course of the discussion, C&L (2005, 61) appear to suggest that the
762 reference to ‘‘brute intuitions’’ is in one or the other way illegitimate. They ask their
763 opponents for an argument in favour of MC, ‘‘beyond just appeal to brute intuitions’’
764 (my emphasis). They do not, however, make any suggestion what the problem with
765 the reference to intuition might be. And in fact they cannot be expected to have any
EC
766 general concern about it, for they do employ reference to intuitions for their own
767 argumentative purposes in other places.27 Thus, we can assume that C&L have no
768 general reservations against arguments in terms of intuition, after all.
769 But let us now turn to C&L’s third argument against Incompleteness Arguments.
RR
770 Over and above their SSAs, they try to rebut Incompleteness Arguments with ref-
771 erence to the (alleged) fact that they are metaphysical arguments. ‘‘Moderate and
772 Radical Contextualists who use such arguments’’, they (2005, 11) conjecture, ‘‘are
773 typically deeply confused about the relationship between semantics and meta-
774 physics’’. ‘‘These arguments are not about language’’, they contend, but ‘‘they are
CO
775 about various non-linguistic aspects of the world’’—thus, instead of being seman-
776 tical arguments, they are metaphysical arguments, they argue.
777 This diagnosis may strike us as a bit surprising. They establish it as follows.
778 Referring to an argument we dealt with above, they (2005, 61) start by saying that
779 ‘‘[a]ccording to Bach, [the sentence ‘‘Steel isn’t strong enough’’] doesn’t express
UN
780 genuine truth conditions’’. This is indeed the claim we know from Bach’s (1994b,
781 127) article, and at least at first sight it does not appear very much as a metaphysical
782 claim. But C&L continue by giving three reformulations of the claim. The first
783 reformulation is ‘‘There’s no proposition of the sort that steel isn’t strong enough’’.
27
FL01 See, e.g., C&L (2005, 1, 62 n1, 151).
123
784 This already sounds a bit like a metaphysical claim. The second reformulation is
785 ‘‘The italicized expression in the previous sentence just doesn’t refer to something
786 truth-evaluable’’. The third reformulation of Bach’s argument is ‘‘There’s no such
787 thing as steel not being strong enough simpliciter’’. Now this third reformulation,
Author Proof
788 indeed, heavily sounds like a metaphysical claim, namely, a statement about the
789 question whether a certain kind of property (being strong enough simpliciter) exists
F
790 or not. Consider also the following description they give of Incompleteness Argu-
791 ments:
OO
792 Incompleteness Arguments are always simple (so simple that they might not
793 deserve the moniker ‘argument’). Typically, all they amount to is a claim like
794 the following:
795 Consider the alleged proposition that P that some sentence S semantically
PR
796 expresses. Intuitively, the world can’t just be P simpliciter. The world is
797 neither P nor not P. There’s no such thing as P’s being the case simpliciter.
798 And so, there is no such proposition. (C&L 2005, 11)
799 In this passage, the Incompleteness Argument appears as a statement concerning
800 the question whether a proposition such as ‘that P is the case’ (simpliciter) exists.
801
D
Again, this sounds like a metaphysical statement (namely, concerning the existence
802 of certain propositions).28
TE
803 Now the first question we may ask ourselves is, even if the Incompleteness
804 Argument really were a metaphysical argument, why should this be considered as
805 problematic? According to C&L, the reason is twofold. Firstly, Incompleteness
806 Arguments would turn out to be inconclusive; for ‘‘even if they were good argu-
807 ments, [. . .] no semantic conclusions [would] follow from these arguments even if
EC
808 they were sound’’. Secondly, C&L claim that ‘‘considered [. . .] as metaphysical
809 arguments, they are unsound’’ (2005, 11).
810 Apparently, C&L assume that if an argument is a metaphysical argument then it
811 cannot possibly be used to support semantical conclusions. The first problem with
RR
812 their argument is then, however, that this assumption actually is not very plausible.
813 We can, for instance, use a meteorological premise, such as the assumption that
814 tomorrow there will be a hurricane in Barcelone, in order to support a football
815 conclusion, such as that Barça will fail to get in the lead in the Primera División
816 (because their match will have to be cancelled). Indeed, it also appears that we can,
CO
817 for example, use the metaphysical premise that there can be no non-descriptive
818 meaning to support the semantical conclusion that the meaning of interrogative
819 sentences cannot be non-descriptive. Thus, while I cannot see any reason why we
820 should endorse the premiss C&L seem to employ, that metaphysical arguments
821 cannot support semantical conclusions, we have reason to avoid this assumption. If
UN
822 we avoid it, however, then C&L’s argument appears to break down even before
823 starting.
28
FL01 The sketch C&L draw here of Incompleteness Arguments is inconsistent concerning the question what
FL02 kind of metaphysical issues Incompleteness Arguments are supposed to answer: either the existence of
FL03 certain propositions, or else the existence of certain properties.
123
F. C. Doerge
824 Anyway, the argument fails for a second reason. Let me begin with a comment
825 on the paraphrases C&L provide of Bach’s argument. If I state, about the coffee on
826 my desk, that it is not very hot, then this may be paraphrased approximately as the
827 claim, about a possible very hot coffee standing on my desk, that it does not exist,
Author Proof
828 and about a colder one standing on my desk, that it does exist. Stated this way, my
829 complaint may appear as a metaphysical claim. Probably, it would be possible to
F
830 paraphrase virtually every statement in a similar way, replacing ascriptions of
831 property with metaphysically-sounding double ascriptions of existence. However,
OO
832 this does certainly not justify the conclusion that the original claim is a metaphysical
833 one. My remark about coffee, for example, is rather not (as its author I am in a good
834 position to tell). If C&L’s test says otherwise, then this is an indication that the test
835 is not reliable.
836 Keeping this in mind, let us turn to C&L’s favourite example. Bach (1994a, 269)
PR
837 says about sentences such as ‘‘Steel isn’t strong enough’’ that ‘‘though syntactically
838 well formed, [they] are semantically or conceptually incomplete, in the sense that
839 something must be added to the sentence for it to express a complete and deter-
840 minate proposition’’. As we saw, C&L provide several paraphrases of this argument,
841 such as ‘‘There’s no proposition of the sort that steel isn’t strong enough’’, and
842 ‘‘There’s no such thing as steel being not strong enough simpliciter’’, which sound
D
843 like metaphysical statements (concerning the existence of propositions in the one
844 case, and of facts in the other). These paraphrases, however, are indeed misleading.
TE
845 What Bach intends to express with his words is not a statement about certain
846 propositions, to the effect that they do not exist. Neither is it a statement about
847 certain possible facts, to the effect that they do not exist. What he aims at, rather, is
848 a statement about certain meanings, to the effect that they are not truth-evaluable,
849 because as they stand they are somehow incomplete. In particular, Bach does not, as
EC
850 one of C&L’s paraphrases insinuates, deal with the question whether ‘that steel isn’t
851 strong enough’ exists or not. Instead, he is concerned with the quite different
852 question whether ‘that steel isn’t strong enough’ is truth-evaluable or not. Stated in
853 an unpretentious way, this question does not even sound like a metaphysical one. It
RR
854 is reasonable enough, I think, to assume that this defence applies to other Incom-
855 pleteness Arguments as well. C&L’s argument fails, because the surprising premise
856 it relies on is unsound: Incompleteness Arguments are by no means metaphysical
857 arguments.
CO
858 6 Summary
859 Let me summarise the findings of this paper. We were examining the central
860 argument of IS, an argument ‘by elimination’, which says that there is no acceptable
UN
861 alternative to C&L’s very moderate contextualism. They try to convince us of this
862 view by assuming that there are exactly two alternatives to their account, MC and
863 RC, and by arguing that RC is inconsistent and absurd, while MC collapses into
864 RC. The alleged reason for this collapse is that each and every argument in favour
865 of either context sensitivity or semantical incompleteness commits one to the
866 endorsement of RC, because for each such argument a (partial) SSA can be
123
867 construed, showing that the application of this argument cannot be restricted to any
868 finite number of sentences, but will extend to literally each and every sentence.
869 Additionally, they suggest that Incompleteness Arguments use metaphysical
870 assumptions for conclusions about language, which on their view makes them.
Author Proof
871 As our discussion showed, however, the arguments they give are not convincing.
872 First of all, it is not true that MC and RC are the only alternatives to their account.
F
873 The assumption on which this view apparently rests, that context sensitivity and
874 semantical incompleteness are either the same thing or at least coextensional, is
OO
875 unfounded. Building on the assumption that the identity of context sensitivity and
876 semantical incompleteness cannot just be taken for granted, I suggested adopting the
877 view that there is no context sensitivity (‘‘Real SM’’, as I call it), together with a
878 moderate view of the extension of semantical incompleteness. Among the advan-
879 tages this view has is the fact that by completely rejecting the context sensitivity of
PR
880 semantic content it is more obviously consistent with the Basic Idea of C&L’s
881 account than their own theory. Assuming this account, the SSAs against context
882 sensitivity can be ignored, while the SSAs against Incompleteness Arguments
883 remain significant.
884 These, however, turn out to be unconvincing, for several reasons. C&L take two
885 kinds of Incompleteness Arguments into consideration, Open Question test argu-
D
886 ments and Felt inability test arguments. The success of their SSA depends on the
887 assumption that the two arguments exhaust the range of possible Incompleteness
TE
888 Arguments. They really do not, however, and this is one reason why the SSA against
889 Incompleteness Arguments fails. Furthermore, the SSAs which C&L direct against
890 those two Incompleteness Arguments are unsuccessful, too. It is possible to employ
891 Open Questions and Felt Inabilities as indications of semantical incompleteness
892 without being commited that all sentences are semantically incomplete, because
EC
893 neither do Travis-questions indicate the relevant kind of semantical incompleteness,

894 nor is it the case that all questions provoke Travis-questions. It does not seem that
895 they are able to keep their promise, to construe Travis-questions for each and every
896 sentence.
RR
897 Finally, Incompleteness Arguments do not fail by virtue of being metaphysical

898 arguments for a semantical conclusion. First of all, they appear to assume that
899 metaphysical arguments cannot support semantical conclusions; but they do not
900 explain why we should believe such a thing. Furthermore, the interpretation of
901 Incompleteness Arguments as metaphysical claims, which they assume, rests on
CO
902 several misleading formulations they give of these arguments. Insofar as these
903 formulations do present Incompleteness Arguments as metaphysical statements,
904 they misrepresent what the defenders of Incompleteness Arguments have in mind;
905 thus, there is no reason why the adherent of Incompleteness Arguments should
906 subscribe to these inappropriate ‘versions’ of ‘her’ claim.
UN
907 So the central aim of IS, the attempt to prove the truth of their own account by
908 eliminating all other accounts, turns out to collapse for quite a number of reasons.
909 Their view of the range of competing accounts is too narrow,29 the structure of their
910 argument is inconclusive, and the premisses on which their argument rests are
29
FL01 There are additional types of views which their account either ignores or overlooks; see.
123
F. C. Doerge
911 unsound. Thus the claim they make, that their own theory remains as the only game
912 in town at the end of the day, turns out to be too optimistic after all.
913 Acknowledgements I should like to thank, Ernie Lepore and for their valuable hints and comments on
Author Proof
914 this and an earlier draft. Work on the issues dealt with in this paper was supported by.
F
915 References
OO
916 Bach, K. (1994a). Semantic slack: What is said and more. In S. Tsohatzidis (Ed.), Foundations of speech
917 act theory (pp. 267–291). London: Routledge.
918 Bach, K. (1994b). Conversational impliciture. Mind & Language, 9(2), 124–162.
919 Bach, K. (2006). The excluded middle: Semantic minimalism without minimal propositions. Philosophy
920 and Phenomenological Research, 73(2), 425–442.
921 Borg, E. (2007). Minimalism versus contextualism in semantics. In G. Preyer & G. Peter (Eds.), Context-
PR
922 sensitivity and semantic minimalism (pp. 339–359). Oxford: Oxford University Press.
923 Cappelen, H., & Lepore, E. (1997). On an alleged connection between indirect speech and the theory of
924 meaning. Mind & Language, 12(3–4), 278–296.
925 Cappelen, H., & Lepore, E. (2005). Insensitive semantics: A defence of semantic minimalism and speech
926 act pluralism. Oxford: Blackwell.
927 Carston, R. (1988). Implicature, explicature, and truth-theoretic semantics. In R. Kempson (Ed.), Mental
928 representation: The interface between language and reality (pp. 155–181). Cambridge: Cambridge
929
930
D
University Press.
Carston, R. (2008). Linguistic communication and the semantics/pragmatics distinction. Synthese, 165,
931 321–345.
932 Crimmins, M. (1992). Talk about beliefs. Cambridge, MA: MIT Press.
TE
933 Kaplan, D. (1989). Demonstratives. In J. Almog et al. (Eds.), Themes from Kaplan (pp. 481–563), Oxford:
934 Oxford University Press.
935 Korta, K., & Perry, J. (2007). Radical minimalism, moderate contextualism. In G. Preyer & G. Peter
936 (Eds.), Context-sensitivity and semantic minimalism (pp. 94–111). Oxford: Oxford University Press.
937 Perry, J. (Ed.). (1993). Thought without representation. In The Problem of the essential indexical and
EC
938 other essays (pp. 205–225). Oxford: Oxford University Press.

939 Récanati, F. (1993). Direct reference: From language to thought. Oxford: Blackwell.
940 Sperber, D., & Wilson, D. (1986). Relevance. Oxford: Blackwell.
941 Taylor, K. (2001). Sex, breakfast, and descriptus interruptus. Synthese, 128(1–2), 45–61.
942 Travis, C. (1985). On what is strictly speaking true. Canadian Journal of Philosophy, 15(2), 187–229.
RR
CO
UN
123
View publication stats MS Code : LING422 h CP h DISK

The Collapse of Insensitive Semantics

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

The Collapse of Insensitive Semantics

Загружено:

Авторское право:

Доступные форматы

See

The collapse of insensitive semantics

Article in Linguistics and Philosophy · April 2010

Friedrich Christoph Doerge

Nature of (Linguistic) Meaning View project

'What is an "illocutionary act"?' View project

The user has requested enhancement of the downloaded file.

Here are the proofs of your article.

• You can submit your corrections online, via e-mail or by fax.

• The printed version will follow in a forthcoming issue.

Author Query Form

Query Details required Author’s response

The collapse of insensitive semantics

10 semantical incompleteness and context sensitivity, they overlook some possible

15 contention that arguments in favour of semantical incompleteness employ meta-

19 Keywords Context sensitivity Insensitive Semantics Semantic content

49 incomplete.5 So in the main,6 C&L construe the difference between MC and RC as a

53 in C&L’s view RC is a completely unacceptable view. In order to convince us of

75 suggesting that the sentence under consideration is context-sensitive. Incompleteness

79 employ an Incompletess Argument, or a Context Shifting Argument, in either case do

117 incomplete qua semantic entity’’ (2005, 7).

198 2 General analysis of C&L’s argument by elimination

206 And so, there is no such proposition. (C&L 2005, 11)

209 expressed by some sentence S is no more than a propositional fragment’’. That a

285 clear definition of ‘semantical (in)completeness’. That a sentence is incomplete,

317 (4) Some sentences trigger a Felt Inability.

327 3 Analysis and defence of the ‘Open Question test’ argument

337 This sentence, according to Bach (1994b, 127), is ‘‘semantically or conceptually

451 argument is based:

454 This assumption is challenged in the following way:

456 (8) All sentences are subject to Travis Questions.

457 (9) Travis-questions are a variety of Open Questions.

ness, (2) implies RC/commits one to RC.

482 sentence to be truth-evaluable. Accordingly, it is quite pointless, as far as the truth-

487 of the relevant variety ‘‘incompleteness indicating questions’’ (IIQs).

492 Questions employed by some defenders of MC.

500 every sentence whatsoever (2005, 67).

571 4 Analysis and defence of the ‘Felt Inability test’ argument

578 The ‘‘Felt Inability test’’ argument:

641 of MC reacts to each Travis-question by construing a sentence which is less

727 the effect that if we take indexical expressions as an indication of semantical

893 neither do Travis-questions indicate the relevant kind of semantical incompleteness,

897 Finally, Incompleteness Arguments do not fail by virtue of being metaphysical

938 other essays (pp. 205–225). Oxford: Oxford University Press.

Вам также может понравиться