Jaszczolt, K. M. (2000) The Pragmatics of Propositional Attitude
Reports, Elsevier, 218 pp.
Edward Garrett, University of Virginia
Volume 4 of the CRiSPI (Current Research in the
Semantics/Pragmatics Interface) series, this book is a specially
commissioned collection of papers on the subject of
propositional attitude reports. Although the contributions
come from a variety of perspectives, the book is likely to
interest philosophers and pragmaticians more than general
linguists.
The volume begins with a short introduction by the editor.
Jaszczolt briefly surveys recent work on the subject, and then
summarizes the authors' contributions. Representative of the
problem of propositional attitudes is the problem of opacity.
That is, why can coextensive terms be substituted without
affecting truth conditions in matrix clauses (1), while
embedded substitutions appear to have dramatic effects
on truth conditions (2)?
(1) Superman/Clark Kent can fly.
(2) Lois Lane believes that Superman/Clark Kent can fly.
Jaszczolt promises that the different angles from which this
and related problems are approached will constitute the
strength of the volume.
In Chapter 2, S. Schiffer discusses and rejects the "hidden-
indexical" theory of propositional attitude reports. This
theory takes attitudes like belief to be three place relations
between a believer, a proposition believed, and a mode
of presentation (MoP) under which the proposition is believed
(the so-called 'way of 'belief'). Schiffer argues that the hidden-
indexical theory falls to several objections, including a logical
form problem, an application to speech-acts problem, and a
meaning-intention problem. In its place, he proposes an
"indexical" theory inspired by F. Recanati's work. The
indexical theory holds that 'believe' is a two place relation.
Instead of believing propositions under MoPs, subjects
believe "quasi-singular" or "quasi-quasi-singular" propositions,
which are propositions that contain objects and properties
along with modes of presentations of them. Schiffer maintains
that the indexical theory avoids the objections that plague the
hidden-indexical theory.
In Chapter 3, P. Ludlow argues for the applicability of so-called
Interpreted Logical Forms (ILFs) to the theory of propositional
attitudes. By pairing syntactic structure with semantic values,
ILFs individuate beliefs in a more fine grained way than
propositions. Ludlow appeals to psychological principles as
he attempts to explain away Saul Kripke's famous double
Padarewski case:
(3) Peter believes that Padarewski is a pianist.
(4) Peter believes that Padarewski is not a pianist.
where Peter seems to come to believe that Padarewski both is
and isn't a pianist, through being acquainted with Padarewski in
two different ways (for example as the great pianist he sees on
TV, and as the strange man who lives next door).
In Chapter 4, L. Clapp pursues a novel approach to propositional
attitude reports. He notes that previous work has tended to either
deny the intuition of difference in (2), as in the Neo-Russellian
view; or postulate "extraordinary entities" (i.e. MoPs), as in the
hidden-indexical view. Clapp focuses his critique on versions
of the indexical view, where these extraordinary entities are
identified with finely individuated ways of thinking. This kind of
individuation goes against the public nature of language, because
we do not make reference to MoPs when we make propositional
attitude reports. Clapp urges that we instead adopt a "discourse
holistic" view and reject sentential compositionality. Framing his
argument in Discourse Representation Theory (DRT), he then
walks the reader through several simple cases, showing how the
inclusion of contextual information, the postulation of discourse
referents, and rules of anaphora for discourse referents, facilitates
a natural solution to the problem of opacity.
In Chapter 5, M. J. Cresswell sheds new light on Davidson's
paratactic account of 'that'-clauses. He argues that if Davidson's
notion of "samesaying" is understood semantically, then his theory
is not in direct competition with intensional accounts of propositional
attitudes. That is, Davidson does not show that a semantic account
of indirect discourse can get by without appeal to intensional entities.
In Chapters 6 and 7, K. Bach presents an overview of the four
major theories of propositional attitude reports, including (i) the
Fregean view, according to which embedded clauses do not have
their customary references but rather refer to their indirect
references; (ii) the meta-linguistic or quotationalist view, where
belief reports relate believers to sentences (or mental representations)
rather than propositions; (iii) the Neo-Russellian view, which denies
the anti-substitution intuition in (2) above; and (iv) the hidden-
indexical view, which takes belief as a three place relation involving
MoPs. He argues that all of these theories make a common error,
namely their "specification assumption", which is the assumption
that belief reports specify the contents of beliefs rather than
merely describing them. According to Bach, when a person
believes p, what she really does is believe something that requires
the truth of the p rather than believe the proposition p itself.
Furthermore, belief sentences are context-sensitive, and so
substitution sometimes affects truth-conditions, but not always.
Chapters 8 through 10 are pragmatically oriented
discussions of propositional attitude reports. In Chapter 8,
A. Bezuidenhout endorses a relevance-theoretic approach,
and suggests that 'that'-clauses encode "procedural
information", i.e. instructions as to how the expression
is to be processed. She discusses the importance of
pragmatic concepts such as loosening, enrichment, and
transfer, and raises several interesting cases for discussion.
In Chapter 9, the editor suggests that there is a three-way
distinction at the level of belief reports. Along with de re
belief reports, there are also "de dicto 1" and "de dicto
proper" belief reports. Jaszczolt argues that de re is the
default. Finally, in Chapter 10, D. W. Smith tackles
propositional attitude reports from a phenomenological
perspective, and stresses the importance of the
intentionality of mental states for propositional attitudes.
For the most part, the papers were well written and
well organized. However, certain problems remain.
Schiffer's discussion on pp. 18-21 of the "logical form
problem" was somewhat confused. The putative
problem is that the hidden-indexical theory, which takes
belief as a three place relation, should therefore represent
the MoP in the syntax as an argument rather than an adjunct.
Appealing to well-known distinctions in generative syntax
between arguments and adjuncts, Schiffer then proceeds
to show that phrases such as "under what mode of
presentation" behave more like adjuncts than arguments.
The discussion is weak for three reasons: first, I doubt that
advocates of the hidden-indexical theory ever meant for
their "third argument" to be syntactically realizable, and why
Schiffer assumes they should is a mystery; second, he devotes
considerable discussion to an argument by Ludlow against
argument-adjunct asymmetries, rather than citing authoritative
works in syntax which establish the distinction; and third, he
conflates 'wonder' and 'think' together as islands for
wh-extraction, while only 'wonder' is. The editor should
have taken more care with this section.
In a smaller way, I found Ludlow's discussion of the
Padarewski problem to be misleading. In discussing
this and similar cases, Kripke (1979) stressed the belief
itself, and not the language of the belief report. The
question is: what does Peter believe? Echoing Kripke,
"Does he or does he not believe that Padarewski is a pianist?"
In contrast, Ludlow's discussion revolves around the
language of the belief report. He suggests that either
there are two Padarewskis in our lexicon, Padarewski 1
and Padarewski 2, or there is just no situation in which
someone would report both (3) and (4), i.e. it is
linguistically implausible that these reports would occur
together in the same situation. But this is not the issue:
the issue is what Peter believes.
I was surprised at the overall lack of argumentation
accompanying several authors' rejection of Salmon's (1986)
Neo-Russellian view of propositional attitude reports.
As mentioned, the Neo-Russellian view denies that the
contrast in (2) is truth conditional. Schiffer, Clapp, and
Bach all reject this view as obviously counterintuitive.
Bach approvingly cites Richard (1990), who writes that it
would take "bribery, threats, hypnosis, or the like" to get
people to think that the two versions of (2) have Lois
believing the same thing. Yet elsewhere, Bach frequently
appeals to other supposedly infallible intuitions. For
example, he claims that we intuit that the two versions
of (1) have the same propositional content, and that the
two versions of (2) "have Lois believing two different
things ... not ... the same thing in two different ways ...
she believes two different things" (p. 107). The problem
is that all of the terms used are theoretically loaded: what
intuitions do people have about propositional contents, or
about the dinstinction between believing two things and
believing one thing in two different ways? If there are
better arguments against the Neo-Russellian view they
should not be hidden behind such unhelpful rhetoric.
I also found Jaszczolt's distinction between de dicto 1
and de dicto proper unclear. To explicate the distinction,
she asks her readers to imagine that Ralph and John are
walking along the beach. Ralph sees a man with a brown
hat, and says "The man in the brown hat is a spy". John
later reports, "Ralph believes that Ortcutt is a spy". The
de dicto 1 reading is opaque: Ralph thinks, unknown to
John, that the man is Smith and not Orcutt; de dicto 1
involves a referential mistake. In contrast, in the de dicto
proper situation John has no idea who the man in the brown
hat is. I do not see the difference: in both situations John
has a belief about 'that man', i.e. the man he sees with a
brown hat. There is a question perhaps about whether he
already had the belief (in the case of believing something
about Smith, who he may have already known), and
perhaps a question about whether or not the 'brown hat'
was crucial to the belief, but in Jaszczolt's hypothetical
scenario the readings appear to be identical. (In fact so
is the de re report.) She should explicate the distinction
more clearly and further justify its importance.
Overall, the collection was quite interesting. I do not
think, however, that readers will benefit much from the
papers appearing together as a book rather than separately
in various specialty journals. The mandate of the CRiSPI
series is to remove confusion and lack of purpose in
literature on the semantics/pragmatics border, and to
take steps towards clarity and cautious consensus.
This volume did not reach any consensus, and the
papers were not as closely intertwined as the editor
may have hoped.
Bach's Chapter 6 could have served as a great introduction
to the volume, since it was very nicely written and it presented
the four major views of propositional attitude reports in a
concise and interesting way. Had this been the case, several
papers, including Cresswell's, Bezuidenhout's, and Smith's,
would not have fit in. But had the volume been restricted to
the more philosophical pieces, it would have had greater
cohesion and would probably have appealed to a larger
audience, including students.
Also, if the papers had appeared in journals instead of this
book they might have reached us earlier. The vast majority
of citations come from 1997 or before, with only a few from
1998 and 1999. The only previously published piece, Bach's
Chapter 7, appeared in "The Pacific Philosophical Quarterly"
as early as 1997, and this article served as a sequel to his
contribution in Chapter 6! For whatever reason, the book's
journey to publication was not swift.
Kripke, S. 1979. A puzzle about belief. In A. Margalit, ed.,
Meaning and use. Reidel: Dordrecht.
Salmon, N. 1986. Frege's puzzle. MIT: Cambridge.
Richard, M. 1990. Propositional attitudes. Cambridge U:
Cambridge.
Edward Garrett has just finished his Doctoral thesis from
the University of California at Los Angeles on
"Evidentiality and assertion in Tibetan." His research
focuses on semantics, philosophy of language, and
Tibeto-Burman linguistics.
|
