You are on page 1of 8

Journal of Linguistics

http://journals.cambridge.org/LIN
Additional services for Journal

of Linguistics:

Email alerts: Click here


Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here

S. Kuno, Functional syntax: anaphora,


discourse and empathy. Chicago: University of
Chicago Press, 1986. Pp. 320
Susan Wright
Journal of Linguistics / Volume 24 / Issue 02 / September 1988, pp 553 - 559
DOI: 10.1017/S0022226700011968, Published online: 28 November 2008

Link to this article: http://journals.cambridge.org/


abstract_S0022226700011968
How to cite this article:
Susan Wright (1988). Journal of Linguistics, 24, pp 553-559 doi:10.1017/
S0022226700011968
Request Permissions : Click here

Downloaded from http://journals.cambridge.org/LIN, IP address: 163.200.81.46 on 11 Feb 2015

REVIEWS

take to be such), it should be stressed that the mistakes are not really his.
Rather, they are inherent in Chomskyan linguistics, and Pateman deserves
credit for having spelled out several of its aspects that usually remain
implicit. Although most of Pateman's conclusions are predictable, given his
general stance, one often has the feeling that he engages in genuine argument,
and not only in denunciation. In this he is helped by the fact that, unlike the
great majority of his fellow Chomskyans, he has read widely and actually
knows the views he wishes to refute.1
[i] I wish to thank Trevor Pateman for his comments on the first version of this review.

REFERENCES
Andersen, H. (1973). Abductive and deductive change. Lg 49. 765-793.
Bhaskar, R. (1979). The possibility of naturalism. Brighton: Harvester Press.
Chomsky, N. (1986). Knowledge of language. New York: Praeger.
Itkonen, E. (1978). Grammatical theory and metascience. Amsterdam: Benjamins.
Itkonen, E. (1983a). Review of J. Katz, Language and other abstract objects. Lingua 60.
238-244.
Itkonen, E. (1983b). Causality in linguistic theory. London: Croom Helm; Bloomington:
Indiana University Press.
Itkonen, E. (1984). Concerning the ontological question in linguistics. Language and
Communication 4. 241-246.
Lass, R. (1980). On explaining language change. Cambridge: Cambridge University Press.
Lewis, D. (1969). Convention. Cambridge, MA: Harvard University Press.
Pateman, T. (1983). What is a language? Language and Communication 3. 101-127.
Ringen, J. (1977). On evaluating data concerning linguistic intuition. In Eckman, F. (ed.),
Current themes in linguistics. New York: John Wiley.
Saunders, J. & Henze, D. (1967). The private-language problem. New York: Random House.
Reviewed by ESA ITKONEN,

Department of General Linguistics,


University of Turku,
Finland.
(Received 20 October 1987)

S. Kuno, Functional syntax: anaphora, discourse and empathy. Chicago:

University of Chicago Press, 1986. Pp. 320.


Kuno collects, elaborates and distils ideas, arguments and principles which
provide the basis of a series of earlier papers on different aspects of what he
calls functional syntax. The major parts of this approach to traditional
syntactic problems are the direct discourse perspective and the empathy
perspective. His main concern is to show that an approach to syntax which
has the position, attitudes and perceptions of the speaker at its centre, may
be used to account for syntactic phenomena which prove problematic for
mainstream syntactic theory. The work is strongly data-based, and deals
with the problems of anaphora and coreference in sentences with pronouns
and reflexives - the same data which provide the basis of the constant
development of the research programme that is the generative enterprise.
19

553

LIN 24

JOURNAL OF LINGUISTICS

Kuno's motive in focusing on topics like anaphora and coreference is in


part ideological. His aim is to demonstrate that notions which have been
considered to be non-syntactic (belonging to the area of performance rather
than competence and thus not appropriate in dealing with syntactic facts),
can be described and rigorously applied in syntactic analysis. So he reviews
the arguments, conclusions and failures in successive accounts within the
TG tradition from Ross to GB theory. This amounts to a challenge to the
mainstream thinking, using the strategy of one trained in the school. In his
introduction, Kuno asserts that in principle at least, a functional perspective
should not be incompatible with, say, GB theory. He attempts to show this
in practice by reformulating GB definitions to incorporate conclusions of a
functionalist critique of the theory's ability to handle coreference facts in
language.
Among the assumptions attacked and subsequently revised is c-command:
specifically, on the basis that it cannot sufficiently account for anaphoric
binding. Kuno injects a functional perspective into GB by proposing the
indefinite expansion of the domain of anaphor binding. Expansion rests on
the assumption that NP's can be bound by any NP's in any S's provided that
a certain ' command' relationship is met and that a set of semantico-syntactic
conditions if fulfilled (75). This command relationship is a 'chain-ofcommand' principle (based on Langacker, 1969) interpreted as a semanticosyntactic constraint which is not absolute. If coindexing of anaphors with
commanding NPs is free, then the chain-of-command principle is used to
filter out sentences in which a weak controller has been chosen over a
stronger potential controller for coindexing. Kuno proposes two scales: a
syntactic one which depends on the syntactic role of the anaphor, and a
semantic one which determines control of an NP for anaphor binding
depending on the semantic (read 'discourse') nature of the NP. Accordingly,
the subject's status as a controller is strongest if the anaphor is a direct object
and weakest if it is a picture noun. In semantic terms, definite animate NPs
are potentially much stronger controllers than definite inanimate NPs,
followed by indefinite unspecific pronouns (like nothing, anything, nobody,
anybody) and dummy it.
The direct discourse perspective and the empathy perspective form the
heart of Kuno's functional approach. He also mentions two other wellestablished areas in discourse approaches to syntax: perceptual strategies
and the flow of information in sentences in discourse. The former involves
using performance (presumably pragmatic) factors in explaining phenomena,
the latter concerns the role of topic, comment, presupposition, given and new
information in sentence and discourse structure.
The direct discourse perspective is applied in the analysis of pronominalization and coindexing of NPs in complement clauses of direct
discourse verbs like say, tell, ask, complain, worry, bother, and so on. The
roots of the perspective are the performative treatments current in the 1960s
554

REVIEWS

by Ross, Lakoff and others. Following from the argument that the
complements of these verbs have a direct discourse representation as their
underlying structure, Kuno suggests that the pronominalization and
coreference facts are best accounted for by marking the (discourse) role of
relevant NP arguments in the sentence. That is, the subject, direct object,
indirect object, etc. of direct discourse verbs are marked with respect to
whether 'they represent the speaker or the experiencer of the words or
thoughts represented by the complement clause, and in the case that the
verbs involve actual speech, with respect to whether they mark the addressee
of the utterance' (108). He uses the term LOGOPHORIC to refer to these verbs,
and formulates a logophoric pronoun rule which states that a full (nonpronominal non-reflexive) NP in the complement clause of a logophoric verb
cannot be coindexed with the logophoric NPs in the main clause. This rule
then predicts the marginality or unacceptability of (i):
(1) That Ali, was the greatest boxer in the world was repeatedly claimed
by him,.
[He t + logo] claimed repeatedly [Ali was the best boxer in the
world]
It also accounts for the acceptability of the following sentence where the
main clause NP is coindexed with a full NP in the complement:
(2) Those who trusted John, were betrayed by him, repeatedly.
[He betrayed those who trusted John repeatedly]
The verb betrayed is not used as a logophoric verb here. Kuno goes on to
apply this principle to pronominalization facts in that-cMled clauses with
emphatic reflexives and picture noun reflexives. He argues that problems of
pronominalization are often semantic or pragmatic in nature. For instance,
the sentences in (3) have identical phase structure configurations but still
contrast:
(3) (a) ?? That Mary had the nerve to call John, crazy to his, face
bothered him,.
(b) ? That Mary had the nerve to call John, crazy behind his, back
bothered him,.
(c) ?/?? That Mary had the nerve to call John, crazy in front of
his, friends bothered him,.
(d) That Mary had described John, as crazy in front of his, friends
bothered him,.
In these sentences, the contrast has to do with how difficult or easy it is for
the speaker to represent a third party's internal feelings as if in quotation
marks (i.e. as logophoric complements vis-a-vis the experiencer NP) or to
represent it as his (i.e. the speaker's) own knowledge (i.e. as non-logophoric
complements). The logophoric character of a verb and its complement
555

19-2

JOURNAL OF LINGUISTICS

arguments provides the basis for a series of constraints which are formal
statements of intuitions of acceptability, marginality or ungrammaticality
that are hard to account for using purely syntactic principles of command.
The most interesting feature of Kuno's approach is the empathy
perspective. This refers to the fact that propositions are encoded in syntactic
structures which are egocentric. Put another way, the syntactic encoding of
propositions is oriented in terms of the speaker's point of view or 'cameraangle' of the events, people and places commented on. The notion of point
of view has been dealt with in mainstream pragmatics and semantics in
treatments of the deictic categories of tense, person and place, and in
linguistic stylistics. But Kuno's study represents a radical application of the
argument (used by Benveniste (1966), for instance) that speakers/locutionary
subjectivism is grammaticalised in the syntax of a language. The basic
motivation for applying the empathy analysis to English sentence syntax
comes from Japanese, in which camera-angles are systematically encoded in
the pronominal and verbal systems. In particular, Japanese has two series of
' giving' verbs, used as ' supporting verbs to represent the speaker's empathy
with the referent to the subject/nonsubject NP in nonstative sentences' (245).
Essentially, the verbs are subject to an empathy constraint which dictates the
relationship between the subject and indirect object thus (246).
(4) (a) Yar-u requires that E(subject) > E(dative)
(b) Kure-ru requires that E(subject) < E(dative)
So, while the following sentences are similar in that they both mean 'Taroo
gives money to Hanako' and they both imply that the act described is
beneficial to the recipient, they contrast with respect to whose point of view
the speaker is representing:
(5) (a) Taroo ga Hanakoni okane o yar-u Taroo > Hanako
to money give
(b) Taroo ga Hanakoni okane o kure-ru Taroo < Hanako
to money give
The essentially egocentric basis of the empathy perspective is nicely captured
in Japanese by the fact that in addition to the (lexical) encoding of point of
view, perspective is also grammaticalized - in the sense that the collocation
of a first person subject with kure-ru yields an unacceptable sentence. Its
unacceptability follows from the principle that the speaker's point of view or
position is necessarily identical to that of the first person pronoun. Kuno
formalizes this principle as the Speech Act Hierarchy, in which the hearer is
placed below the speaker by dictating that the relationship E(speaker) >
E(others) (where > symbolises the camera-angle) be fulfilled. Basically, this
hierarchy captures the intuition that it is difficult for the speaker to describe
an action or state involving him while taking someone else's camera556

REVIEWS

angle. So a verb which contradicts the Speech Act Hierarchy results in


ungrammatical (never mind infelicitous or inappropriate) sentences.
Of course, this contradiction does occur - in English for instance - in
several different ways and guises. Kuno has another empathy hierarchy
which reflects the iconic encoding of empathy relations - the surface
structure hierarchy. This formalizes the basic fact that the subject NP
position is more likely to be the source of the perspective than others. And
these hierarchies all interact. But in English, there is a normative rule which
judges utterances such as the following as inappropriate or impolite:
(6) (a)
(b)

? The journal has accepted a paper by me and Josh,


? I and Josh are going to London on Saturday.

What motivates the reversal of the hierarchy here is a prescriptively imposed


sense of modesty, which prevents the speaker from highlighting his role in the
joint activity talked about. The prominence of the speaker and/or subject is
much clearer, and is unproblematic in terms of potential conflict arising
between consideration for the role of others (or 'modesty') and the empathy
principle in the syntax of the deictic verbs come (up to), go (up to). With these
verbs, as with the representation of relative temporal and spatial positions of
people and events, the centrality of the speaker-subject is grammaticalized.
In English, unlike Japanese, empathy relations are grammaticalized in a
directly iconic way: through the linear word-order. So focus on the
benefactor or beneficiary involves lexical choice (governed by the subcategorization of the verb):
(7) (a)

Josh gave me some money.


received "|
I some money from Josh,
got
)

These sentences illustrate the interaction of the speech act hierarchy and
the surface structure hierarchy. By virtue of its form, the perspective
represented in (7a) is the speaker's, not Josh's. Yet Josh is the subject NP.
There then seems to be a contradiction of the two hierarchies, since in
princip'e, whenever the first-person pronoun is used, the source of the
perspective is the speaker. This contradiction does not occur to the same
degree when individuals other than the speaker are mentioned:
(8) (a)
(b)

Josh gave Agatha some money.


Agatha received some money from Josh.

In these sentences, the speaker's empathy is biased toward the subject NP


in each case. There are mechanisms which enable the speaker to change his
bias, which cohere with Kuno's hierarchies. One is passivization, which alters

557

JOURNAL OF LINGUISTICS

the surface structure, without losing the lexical specificity of the verb
selected:
(9) Agatha was given some money by Josh.
Note though, that in the case of receive, the beneficiary and not the agent is
the unmarked focus. The speaker's perspective could be shifted to the
benefactor by topicalization - which shifts the agent/instrument to initial
position:
(10) From Josh Agatha received some money.
Kuno's assessment of the treatment of anaphora and coindexing in GB is
in part addressed at the level of detail. Often, because the formal analyses
contrast or have a very different basic set of assumptions from Kuno's
functional one, it is useful to have the offending account to hand when
reading his critique. It took me some time to find a particular discussion
referred to as Chomsky, 1982: 88-100, and when I did, I was surprised to find
that its substance was a footnote to the text (actually Chomsky, 1982:

99-100; fn. 24). What started this search was a set of example sentences on
page 171 that had been misprinted and misnumbered - a small but crucial
error.
Kuno's attempt to incorporate the semantic-pragmatic insights of a
functional approach into the formalism of generative grammar is an
ideological exercise (or crusade). Now while it seems plausible to combine the
two at a purely instrumental level - that is, at the level of language-specific
descriptive syntax - there are some problems at the level of the theory of
language that is presupposed.
The fundamentally intra-theoretical position held by Kuno is made very
clear by his attempt to incorporate the logophoric principle in the binding
conditions (which deal with anaphora) in GB theory. It is important to note
here that Kuno's arguments are motivated by intuitions of acceptability (and
marginality) which are not reflected or predicted in GB theory. That is, Kuno
builds his perspective by a process of exemplification and counterexemplification using sentence tokens requiring intuitions about acceptability. The inevitable problem is if acceptability is to a large extent a
matter of pragmatic interpretation, then one can, in principle, apply
pragmatic solutions to their construal and restructuring. Kuno does do this
to the extent that he makes reference to the roles of NP arguments vis-a-vis
their verbs and its complement. In particular, with logophoric verbs, it is the
role/identity of the NPs as speaker/experiencer etc.
However, Kuno's account raises questions concerning the deeper claims of
GB theory about the specific cognitive faculty at the core of linguistic
competence. In particular, a pragmatic approach to anaphora phenomena
challenges the idea that their properties provide the key to the study of
competence. These problems are beginning to be aired in a growing body of
558

REVIEWS

work addressing the interface of pragmatics and grammar (from Haiman


(1985) to Levinson (1987) for instance). In the context of such work, Kuno's
represents the detail and data in this development.
REFERENCES
Benveniste, E. (1966). Problemes de linguistique generate. Paris: Gallimard.
Chomsky, N. (1982). Some concepts and consequences of the theory of government and binding.
Cambridge, MA: MIT Press.
Haiman, J. (1985). Natural syntax: iconicity and erosion. Cambridge: Cambridge University
Press.
Langacker, R. (1969). On pronominalization and the chain of command. In Reibel, D. A. &
Schane, S. A. (eds). Modern studies in English. Englewood Cliffs, NJ: Prentice-Hall.
160-186.
Levinson, S. C. (1987). Pragmatics and the grammar of anaphora. JL 23. 379-434.
Reviewed by SUSAN WRIGHT,

Department of Linguistics,
University of Cambridge.
(Received 10 December 1987)

559

You might also like