I f},...

‘ ..,.w

'35
I .

5,555.5";-

...u ‘r‘v5
- r

4.7.:-

 

. I," ‘1".
H “(-1'~:!":‘l' 535,35:
": .rr"" '4'!

 

 

 

 

ma." 1"?”
“mu”: ... 1
-- ”5”...»

,
1: r
5“,": «.14 "W, ' "r
I w I )1": ,5 1'7 1‘:
5.55.1 . 5-3“ '.

1, urn-7.
,tg'w’h’

.

.. 1- I ..~- ....-, 4

r1" I ,.,Jvm~;I-rw-«
L‘“

I; ‘1“‘L’ . I
I I.“ , ”7.5..."
:_ “51“,“.H-p ,
:,..,,-,., . 4‘:

".23.;
5" ~ ...7.-w'-
.1vI-(II ’,.1 A".
1 .-..
5. 534:. :5

‘ r w
'.l, ..., . 5') .
”Amy

"w'l v l‘l
.5 a ,,
J';‘v):!3~.

“‘52..

. ,
I" 1
.5. - I r . .
1.. . . I”,

. I
15., .

r—u .u. .- ‘4qu
. r“

.41»- ..

45,.“ ‘

h“: u
“AS-m

1,3!
' ., .n. iv
.a

“unr-
1.111
ulugw
‘-.15v"-‘ 5
wu

5.55; .
.'
-H-Iml:‘_ .‘ '
:I‘l 3.: w...“
a.

-5
, ‘
., . n- v (4'7”.‘5.

5.“. '.K W1-
, ”'5... v MU '4

. u v- I Q

, 5 . .

. ;, 13 '.J 5

_ ,j ,

' .- J,‘ '
55-13.:

fw
1:555:55,
w‘ fl-{v‘ r'

.,. -n 1:02!»-
,5, _., _- -;‘1A;n’ 5
5. , ,5. .5: . ‘2

v' -rt; «

 

 

 

 

”5'”.
,u

 

 

:n‘trj 5,5
n1-.

“r: 5'
_,5. .554.
.5 . 7:5:

“L.

I . 5 :5 ”A 5,
’13-“...u7-1 ‘5 r-hn
I" 75‘ ‘l

5 q'
5.: r

.(l r‘
.""_,.'.".'.-a-
cum 1

5 5. .

3m ,.'_'.1—4 .5
,._.,..:.» _r ..

u

--.~ "~-¢r

., _¢ .55

\' «1
.,.. .

m
.d, v;
5,-
ms; '
"'s;_..

5,, , 5 :. ;-- 1:...
"’2‘" r~ arr/I

{.5

H...
n

'1“!

. .53.; . ..

5..
1‘ , '.:,.
v " n. I ""HV" ”'

. , . , . ‘v‘ ‘
"1911“),51 ., » -... ‘ .- ,-:" . :IT. -
. , . "‘11-
.:,5 _;I._,.,,.;. . ,1... _
“Wang-v ',a'- Y

5 5

v3
3.,” _
.5., ,3“..- l

5:». 27:5" ' '
,..:.-_n

m
'“ ,...5::~' “V '
I w

4:124 5
"““L :11».

, n: :5-
,_ 525.4-.IW,‘
J“; L": '

.,

: .‘nwr-y- ---

;.,5...‘.,.;.~
5. ,‘w v \- ‘:_"""\,_.. u r .

u on a .. . .' . ame'T ' ""1"“

,.,:.. mm...- .a f :3: “5...”. a

v1 '1 - w‘

i

..\. . .
'5,.,,4,‘5 ,u»
11‘:-

 

‘vw; .2: v
vii" 1:... ".53.”.-
n— n .vu.

llllllllllllllllllllllllllllllllllllllllllllllllllllllll

1293 00785 0054 F

 

LIBRARY
Michigan $23k;
University

A‘a-n'nhs - mm.-

 

 

This is to certify that the

thesis entitled

Iconic Features of the Syntax of
Wordsworth's Poetry

presented by

Keith w. Slater

has been accepted towards fulfillment
of the requirements for

Master's degree in Linquistics

MW

 

Major professor

Date Auqust 11. 1989

222868 :0

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

 

 

DATE DUE DATE DUE- wDATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
chnS—pd

ICONIC FEATURES OF THE SYNTAX OF WORDSWORTH'S POETRY
BY

Keith William Slater

 

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Linguistics
and Germanic, Slavic, Asian and African Languages

1989

oo4iq@fi

ABSTRACT
ICONIC FEATURES OF THE SYNTAX OF WORDSWORTH'S POETRY
BY

Keith William Slater

This thesis addresses issues of iconic coding in natural
language syntax, within a functional theoretical perspective.
It first presents data arguing for the Viability of iconic
principles as explanations for some observed features of
syntax. Next, an attempt is made to utilize some of the
principles proposed to help explain specific features of the
syntax of William Wordsworth's poetry. In so doing, it seeks
both to provide further evidence in favor of the existence of
iconic motivations in human language and to show that
identifiable syntactic principles contribute to some of the
impressionistic effects of Wordsworth's art. Finally, some
initial implications of this investigation for description of
the contribution of purely linguistic phenomena to literary

works are discussed.

 

give

debt

ACKNOWLEDGMENTS

For the successful completion of this research I wish to
honor and praise to God and to acknowledge my immense
to the following persons:

Drs. David Lockwood, Carolyn Harford and Ruth Brend,
whose valuable comments guided the completion of this
work

Dr. Thomas Payne and Mr. Lynn Conver, who first
introduced me to the idea of iconicity in language

Mr. Tim Pulju and Mr. Fraser Bennett, who have inspired
me to pursue linguistic investigations and have reminded
me of how much more there is to learn

Mr. David Vanamburg, who provided valuable assistance in
preparing the manuscript

My parents, who have been constant sources of love and
encouragement throughout my program and my life, and

The members of Graduate InterVarsity Christian
Fellowship at MSU, who have helped me to remember that
all other knowledge is secondary to knowing God.

TABLE OF CONTENTS

List of Tables ................................ .... ....... ...v
1. Introduction and Goals ............... ..... .......... ....2
1.0 Introduction .............. ............. ....... ......2
1.1 Theoretical Framework ....... . ...... ...... ....... ....4
1.2 Iconic Coding ...... ... ...... ........ ........... .....6
1.2.1 Replication................... ............. ...7
1.2.2 Images and Diagrams............... ........ ....8
1.2.3 The Coded Domain ......... ............... ...... 9
1.3 Iconicity and Levels of Language....................9
1.4 Iconicity as a Cognitive Principle ............... ..11
1.5 The Analysis of Written Language..... .............. 13
1. 6 Limitations of this Thesis............... .......... 15
2. Ar rguments for Iconicity............ ...... ......... ..... 17
2.0 Introduction ...... . ..... ... ......... ...............17
2.1 Sequential Ordering.......... ................... 17
2. 2 Similar Form Corresponds to Similar Function.......l8
2. 3 Linguistic and Conceptual Distance.................20
2. 4 Markedness.. ........................ ...............24
2.5 The Speech of Children ..................... ........25
2.6 Developments over Diachrony... ...... ... ...... ......29
2.7 Processing Strategies......... ........ .............32
2. 8 The Lexicon.... ............ . ........ .... ...... .....35
2.9 Discussion.. ..... ........... ............ ...... ..... 36
3. An Analysis of Iconicity
in Wordsworth' s Poetic Syntax ........... . ............ ..37
3.0 Introduction. ..... .... ................ .............37
3.1 Nominals ............................ . ...... . ..... ..38
3.2 Preverbal Objects.. ......... .......................49
3.3 Sentences.... .......................... ............57
3.3.1 Punctuation...... ..... . ..... .................61
3.3.2 Subordination... ...... .......................63
3.3.3 Apposition ............................... ....64
3.3.4 Parallelism .......... .. ..... .. ......... ......65
3. 3. 5 Ellipsis. ...... .......... ..... ...............67
3. 3. 6 Discussion ........... ........................71
4. Summary and Implications................... ............ 76
4.0 Summary .................. ............... ........... 76
4.1 General Comments..... ......................... .....78
4.2 Syntax and Meaning ...... ..... ...................... 79
Appendix: Notes on the Corpus ..... . ..... ..... ........ ......83
Bibliography ...... ... ............. . ...... .. ................ 84

iv

TABLE

II

III

IV

VI

LIST OF TABLES

PAGE

Referential Distances in a Meditative Section

of Wordsworth' 5 Work .......... ...............42
Referential Distances for Written English

Narrative..... .......... .... ................. .43
Referential Distances for Full and Reduced

Nominals ...... ...... ...... ....................44
Referential Distances in Wordsworth' s Poetic

Narrative................... ...... ............46
Punctuation Marks of the Immortality Ode.........62
Levels of Language and Literature... ........... ..80

"One of the reasons linguists have suffered justified
attacks in the past is that we have often assumed that vowel
and consonant counting is the way to show that linguistics is
relevant to literature."

-- Archibald A. Hill (1976:45)

1. INTRODUCTION AND GOALS
1.0 INTRODUCTION

In the development of linguistic investigations, the
description of linguistic systems has often been pursued
without much interest in any potential significance in
discourse of the internal arrangements of those systems.
Although there have been exceptions, such as Tagmemics,
Systemics and others, linguistic theories have generally
considered linguistic systems independent entities, capable
of being described in terms of potential utterances,
essentially without reference to the situational contexts in
which such utterances might occur. It has, for example,
usually been considered beyond the scope of syntactic
investigations to pursue an understanding of any interplay
between the observable constructions in a text and the
structure of the text itself. The grammar of sentences and
the structure of discourse have been treated as essentially
autonomous.

It is possible, however, that there exists some
discoverable set of relationships between the structure of
discourse and the patterns observable at lower levels (e.g.
syntactic) of linguistic coding, if one adopts an approach to
language which admits the possibility that language is but
one of many functions of human cognition, and thus, that
general principles of human cognition may conceivably affect
observable linguistic structures in identifiable ways.

Such an approach, of course, could only be justified by

convincing demonstration of such cognitive influences upon

 

linguistic systems; on the other hand, demonstration requires
prior, purposeful investigation, and the latter has been
largely lacking in theoretical quarters.

This lack has not been complete, however. In
particular, some linguists working within a functional
approach to language have sought to demonstrate effects of
principles of iconicity on natural language syntax, and it is
with this subject that the present study will be concerned.

One specific result of the general lack of interest in
investigating the relationship between text and syntax has
been that linguists have generally had little to say about
literature. Of course, if the linguistic system is truly
autonomous, this should be no great surprise. Language
proper should be considered simply the stuff to be used in
the communication of ideas, but not at all related to the
ideas themselves. Thus, studies of linguistic structures
really should reveal very little about discourse context, and
Hill is wrong in suggesting (as quoted above) that
linguistics ought to contribute more to the study of
literature than simple "vowel and consonant counting"
(1976:45).

If, however, linguistic systems are influenced by the
same sorts of cognitive constraints that help to structure
other human behavior, then we should expect a greater
contribution by linguists to the understanding of literature.

Specifically, we should hope that analysis of linguistic

structures occurring in literary contexts (as well as any
other identifiable contexts) would provide insights into the
way in which particular literary effects are achieved. Such
fairly impressionistic categories as style, register, poetry,
and prose could legitimately be expected to be at least
partially characterizable by reference to specific syntactic
patterns.

The task of this thesis is to undertake just such a
characterization. The idea of syntactic iconicity is
presented and defended in Chapter 2. Thereafter, Chapter 3
attempts to apply the concepts already discussed to a
specific body of literature: the poetic works of William
Wordsworth. This will not only provide further evidence for
the existence of iconic principles in syntax, but also will
demonstrate that some specific effects of Wordsworth's work
can be partially explained with reference to purely syntactic
features found therein. Thus, a better understanding both of

syntax in general and of Wordsworth's poetry will be sought.

1.1 THEORETICAL FRAMEWORK

The theoretical framework for this investigation will be
what is termed a "functional" approach. Such approaches
concern themselves with the relationship between the
observable forms in human linguistic systems and the
functions for which those forms are employed. Language is

not considered an independent entity, to be studied and

 

understood apart from other human behavior; rather, the
assumption is made "that language and its notional/functional
and structural organization is intimately bound up with and
motivated by the structure of human cognition, perception,
and neuro-psychology" (Giv6n 1984:11). Linguistic structures
are seen as an area for investigating the claims of cognitive
psychology regarding the operation of the mind, and the
cross-linguistic generalizations observed by the linguist
are, in turn, expected to help in explaining and/or
predicting the experimentally justifiable claims of the
cognitive psychologist. Such an approach has in fact been
said to View "data of language use, variation, development,
behavior, discourse processing and experimental cognitive
psychology as part and parcel of one empirical complex"
(Givon 1984:10).

In looking for explanations relating to the functions
performed by linguistic systems, theorists must emphasize
motivations for observable structures which are
non-arbitrary. This is necessitated by the desire to
demonstrate that various functions are not haphazardly
performed, but that general cognitive principles motivate the
creation, modification, and/or preservation of structures
which, in essence, are tettet than other logically possible
variations at performing the communicative tasks for which
they are employed, either in a specific language or

cross-linguistically, as the individual case may be. The

better—mess of any one structure over another must then
presumably be defined in terms of human cognitive strategies
and their reflection in the linguistic structures in
question.

Such a view presumes an ability of languages to eeept at
all levels, selecting functionally advantageous structures
over others. Adaptation would most often involve diachronic
change, but could also be relevant to the conscious,
situational choices of a language user, such as we will
explore in Wordsworth's poetry.

The functionalist seeks to explain linguistic structures
both with regard to cross—linguistic generalizations and with
regard to language-specific structures. He attempts to
explain sentence—internal structures with relevant functional
motivations, and then moves on to "the next stage of
syntactic investigation -— the study of texts, and the study
of the functional distribution of various morpho-syntactic
structures within the text" (Giv6n 1984:10-11). It is at
these two levels that the investigations of this work will be

carried out.

1.2 ICONIC CODING

In this study, we will be concerned with iconic coding
in natural languages. This Section will briefly describe
iconicity as a relationship of replication, utilizing images

and diagrams to help in coding some extralinguistic domain.

1.2.1 REPLICATION

Any claim that some particular representational system
operates, in whole or in part, on iconic principles is,
essentially, a claim that "variation in the code replicates
variation in the coded domain" (Payne 1988:4). For the
purpose of this investigation, the "code" to be examined is
the linguistic code, and the replicative behaviors that will
be of interest are generally those which involve
morphosyntactic -— rather than, say, lexical —— phenomena.

This is not to say that iconic principles could only
operate in syntax, however. In fact, if iconicity is
determined to exist at the level of syntax, then a theorist
hoping to present a unified account of linguistic structures
and behavior will naturally expect to find such principles at
other levels, as well. If some motivation is present at all,
then we would hope to find it ubiquitously present, rather
than localized at a particular level.

Taken to a logical extreme, the identification of
iconicity as replication could allow us to modify Payne's
formulation, above, to state: eey variation in the code
replicates variation in the coded domain. Such a step is in
fact taken by Haiman, who defends "a language—learning
strategy which [Haiman] will call the 'isomorphism
hypothesisz' Different forms will always entail a different
communicative function" (1985azl9). At several points in

this thesis, we will invoke this sort of idea in claiming

that certain syntactic constructions, in comparison with one
another, iconically indicate differences in meaning simply by
exhibiting differences in form.

1.2.2 IMAGES AND DIAGRAMS

Pierce (1932) distinguished two types of iconicity:
that which involves some tmeee and that which involves some
sort of diagram. The former, briefly, utilizes a single sign
(a photograph, a statue, an onomatopoeic word) to represent
its referent, while the latter employs a systematic
arrangement of signs in such a way that the relationships
between them (and not the individual signs themselves)
resemble their referent. Examples of iconic diagrams might
include stick figures, stratificational trace diagrams, or
John Madden's CBS Chalkboard elucidations of particularly
interesting football plays. Further discussion of this can
be found in Haiman (1980).

Giv6n (1983) and Haiman (1985a) have suggested that
diagrammatic and imagic iconicity should not be considered
distinct categories, but should rather be viewed as opposite
ends on a continuum of icons. Between the two poles, icons
would be seen as varying in the degree to which they rely on
diagrammatic vs. imagic features, or Vice-versa. Presumably,
constructions could also vary in the extent to which they
could be deemed iconic at all -- an idea which will doubtless
become increasingly meaningful to the reader throughout the

rest of this thesis.

1.2.3 THE CODED DOMAIN

Opinions have differed as to what, exactly, the "domain"
being iconically coded in language may be. Haiman has
claimed that languages utilize iconicity in their
representation of "reality" (1980:537), at least to a great
extent. On this basis, he proposes that universals of syntax
may ultimately be relatable to "properties of the world,
rather than of the mind." This view, however, seems to be
motivated by a rather violent reaction against
transformationalist claims about the mind's mechanisms for
handling language, and it is not strongly supported.

A more cognitively—based View is espoused by Giv6n
(1985:191): linguistic structures mirror our perceptions of
reality. This, of course, allows for some amount of
distortion to take place in the linguistic representation of
external reality, but this seems a rather sensible allowance
to make when dealing with human processing and representation

of the world.

1.3 ICONICITY AND LEVELS OF LANGUAGE

As we have already suggested, iconicity may manifest
itself at various levels of linguistic coding. Giv6n
(1985:189) mentions specifically the "lexical...
propositional... and discourse—pragmatic functional domains"
as potential areas of such manifestations.

Obviously, these domains will be difficult to compare

 

10

with one another, given their relative size and complexity.
Generalizations, however, are possible. Payne (1988z6)
suggests: "as the level of analysis moves from lexical
semantics, through morphology, syntax, and discourse
structure, the nature of the relationships between form and
function tends to become more iconic and less
conventionalized." A natural implication of this stance, and
one which Payne goes on to draw, is that it explains the fact
that languages vary most widely on the lexical level, and
progressively less at higher levels. Thus, while
similarities among lexical items are often strong indications
of genetic relationships among languages, similarities at
higher levels, which are more likely to be due to general
iconic (or other cognitive) principles, are less useful for
drawing such conclusions.

Although this investigation will be primarily concerned
with syntactic structures in its pursuit of iconic
motivations in Wordsworth's poetry, one of the secondary
goals will be to show that iconicity may be found at numerous
levels of linguistic coding. In particular, Chapter 2 will
present several analyses suggesting the presence of iconic
motivations at levels of morphological and suprasentential
(discourse) coding, in addition to the purely syntactic
level. Additionally, Chapter 4 will discuss the significance
of iconicity in terms of how it relates to the conveyance of

semantic information by the structures within lower levels.

 

11

1.4 ICONICITY AS A COGNITIVE PRINCIPLE

Claims such as those advanced in 1.3, above, hinge on
the assumption that iconicity is somehow a general principle
of human cognition; otherwise, its presence in widespread
levels of linguistic coding would be a rather curious thing.
One attempt to describe a cognitive basis for this is found
in Giv6n's (1985) formulation of an "iconicity
meta-principle" which states: "All other things being equal,
a coded experience is easier to store, retrieve and
communicate if it is maximally similar to the experience" (p.
189). This formulation implicitly claims that iconic
principles operate to help meet the cognitive "need to
facilitate processing within real time" (p. 198). In the
definition quoted at the beginning of Section 1.2.1, Payne
also implicitly recognizes cognitive principles at work when
he asserts that the linguistic system (along with those who
manipulate it) is somehow sensitive to replication of
experiences.

A further step along the road down which this line of
thinking leads is taken by Hopper and Thompson (1985).
Examining the categories Noun and Verb cross-linguistically,
they reach the conclusion that "linguistic forms actually
lack categoriality" when viewed simply as forms, outside of
any discourse context. It is only within particular
discourse contexts that realization within a particular

grammatical category (e.g. Noun or Verb) is "IMPOSED ON the

12

form" (p. 179). Here, the authors are arguing for a
"perceptual basis of the cognitive strategies underlying
grammars" (p. 179). Cognitive strategies such as iconicity
would then become links between linguistic form and discourse
function -— which is precisely what other theorists have been
implicitly (or explicitly) claiming.

If iconicity is a general principle of human cognition,
then of course it must be valid for all humans. Further, we
should expect that claims about how cognition works can be
substantiated with supporting data, clearly demonstrating the
effects of whatever mechanism has been proposed.

All of the studies presented in Chapter 2, below, will
be relevant to this issue, since all claim to demonstrate
iconic influences in human language. Three studies, however,
will be of particular interest: Slobin (1985, Section 2.5
below) describes iconic principles in the speech of children;
Greenberg (1985, Section 2.6) deals with iconic motivation
for diachronic linguistic developments; and Prideaux (1987,
Section 2.7) characterizes strategies for information
arrangement. If the claims of these three studies in
particular are found to be valid, they will help in
establishing what Slobin (1985:229) calls "a deeply-rooted
ontogenetic basis" for iconicity in human language, since
they show independent and broad-ranging evidence that
iconicity affects language learning, language change, and the

syntactic presentation of information synchronically. One

13

could hardly accept all (or any) of these claims without then
admitting that some principle of iconicity must be a part of

the human processing ability generally.

1.5 THE ANALYSIS OF WRITTEN LANGUAGE

One area in which some theoretical remarks should be
made is that of synchronic analysis of written, rather than
spoken, language. This is of particular interest, due to the
literary nature of the works which will be examined here. It
must certainly be recognized that a great deal of
intentionality and premeditation (not to mention revision)
lies behind poetry of any literary value; how, then, will our
synchronic linguistic interpretation be affected?

There can be no question that written language generally
exhibits any number of differences from spoken expression.
Planning and revision probably allow for some different types
of organizational structure than does extemporaneous
discourse. Additionally, the lack of interlocutors as active
creators of discourse when it occurs in written form has
obvious implications for the structure of discourse;
turn—taking, for example, is extremely uncommon in

Wordsworth's poetry, which consists mostly of self-revelatory

monologues.
At the level of syntax -- with which we are primarily
concerned —- it is common to find constructions in written

discourse that might be deemed archaic in normal speech; this

14

is traditionally especially true of poetry. Furthermore, the
use of meter and rhyme often limits the choices which a poet
may make, syntactically, so that further syntactic
peculiarities might be expected.

All of these points might serve to convince us that some
phenomena commonly found in spoken language could prove
elusive in written discourse —- and especially so in poetry.
Several considerations, however, may serve to preserve hope
of discovering in Wordsworth's compositions the same sort of
cognitive motivations (specifically, iconicity) as have been
posited for other types of discourse.

First, we should consider the fact that Wordsworth
himself desired to write in a language accessible to the
common man. As we will see in Section 3.0, the poet had in
mind to utilize the everyday expressions of everyday men,
rather than the stylized traditional language of poetry in
his day. Thus, his language is probably lacking some of the
conventions and archaisms often typical of poetry.

Furthermore, in all but a few instances, Wordsworth
chooses not to rhyme in his poetry. One motivation for this
is "to obtain greater syntactic freedom" (Rehder 1981:83).
Again, this will mitigate the amount of deviation in his
poetry from the forms of the spoken language.

More important, perhaps, is the role of iconicity in the
general framework which has been adopted here. We have

accepted iconicity as a cognitive principle, and have

15

proposed that cognitive principles play a major role in
establishing a framework within which human language
operates. If this is truly the case, (as will, it is hoped,
he demonstrated in Chapter 2), then as a matter of course we
will find iconic principles somehow operating at every level,
and in every mode, of linguistic expression. Within this
framework, the real surprise would be to fail to identify
such influences in some area of inquiry.

In fact, given iconicity as a principle which helps to
delineate the structures of human language, we may even hope
to identify areas in which iconic principles are
characteristic of written language; such an area will be
proposed in Section 3.2, where we will see how iconic
syntactic features may help readers to identify grammatical
relations within a sentence exhibiting atypical word order.

The premeditated, artistic nature of Wordsworth's
poetry, then, need not deter our search for iconicity in its

syntax.

1.6 LIMITATIONS OF THIS THESIS

The linguistic study of poetry naturally gives rise to
the desire to provide some sort of definition of the subject:
perhaps even just a partial list of defining characteristics.
Further, a study of iconic features of syntax may lead us to
expect a proposal for describing ell of syntax with

motivations that are either iconic or that derive from some

16

other cognitive pressure.

Neither of these will be attempted here. The first is
simply beyond the scope of this investigation; the second is,
in all probability, impossible. However, general comments
will certainly be made bearing on both issues, insofar as the

current study suggests points relevant to either.

2. ARGUMENTS FOR ICONICITY
2.0 INTRODUCTION

In this Chapter, we will examine data from a variety of
studies, all of which help to illustrate the operation of
iconic principles in linguistic structures. Most of these
examples will be confined to syntactic constructions, but
some will also be included to indicate how iconicity may make

itself felt at other levels as well.

2.1 SEQUENTIAL ORDERING

One of the most obvious areas in which principles of
iconicity may operate is that of describing events in their
original sequence. This is by no means limited to syntactic
constructions, although it may certainly occur there, as well
as elsewhere. In English, for example, the order of the
sentences in (1) will probably be assumed to mirror the

real-life sequence being described:

(1) John hit Harry. Harry hit John.

In order to evoke some other sequence, one would usually have
to indicate it by lexical addition or the use of alternate
syntax: by the addition of ettet or by reversing the order of
the two sentences, for example.

Essentially, this type of iconicity involves a
parallelism between the stream of speech and the flow of
time; both are linear, and thus sequential. Most naturally

in human experience, time moves forward. Likewise, the

17

18

stream of speech advances. It is natural, then, for the mind
to associate the advancing of speech with the advancing of
time, and this yields an iconicity of sequence.

Haiman claims that this sort of iconicity is "by far the
most widespread" (1980:533) among natural languages. He
notes, however, that even this kind of iconicity is not
universal, since a native Burmese speaker, for example, will
interpret instructions given with no grammatical sequence
indicators as simultaneous (c.f. Becker (1975)).

Nonetheless, sequential iconicity is quite common.

Thus, it is likely that no one has ever taken the order of
"Véni, vidi, vici" ("I came, I saw, I conquered") as anything
but an iconic model of the sequential ordering of the

original events described.

2.2 SIMILAR FORM CORRESPONDS TO SIMILAR FUNCTION

Another iconic principle which may be seen to operate
widely within natural languages is that similar form
generally corresponds to similar function. This principle is
one of the fundamental enablers of human language, and its
influences may be found at nearly all levels of linguistic
coding.

Most importantly, this principle helps to account for
systematicity in language at any level. Only on the basis of
an assumption such as this one, for example, can speakers of

a language agree that meaning X will always be indicated by a

19

particular morpheme Y, rather than by any of the extremely
large set of other possible morphemes they could
theoretically choose. This seems perhaps trivial, but it is
a fundamental assumption, necessary to prevent the arbitrary
linguistic sign from being additionally synchronically
inconsistent -- and thus useless for communication. Giv6n
(1984:33) puts it this way: "It is only because the coding
relation between structure and function in syntax is
non—arbitrary, or in some sense iconic, that one could
proceed to infer common function from common structure.

At higher levels, of course, this principle accounts for
the fact that patterning exists in various systems: e.g. that
different verbs exhibit paradigmatic morphological markings
across a range of tense, aspect, number, gender, etc.
categories, and that semantic information can be inferred by
a listener on the basis of, say, the relative positions of
verb and subject in an English interrogative.

The "similar form/similar function" principle is, in
fact, an important basis for human cognition, as has been
demonstrated by, for example, Labov (1973) with regard to
human classificatory behavior. Thus, operations of this
principle in human language may be seen as iconic indicators
of cognitive processing, manifested within the linguistic

system.

 

20

2.3 LINGUISTIC AND CONCEPTUAL DISTANCE

Perhaps the most interesting defense of the existence of
iconic motivations for morphosyntactic phenomena is the
correlation, suggested by Haiman (1983, 1985a) between
conceptual distance and linguistic distance. The former term
is an impressionistic one, used to express how closely
concepts are related to one another in mental representation
-- either conscious or unconscious. Haiman suggests that
greater conceptual separation ("distance") may be iconically
mirrored by greater linguistic separation. A scale of
"diminishing linguistic distance" is employed to help in

determining correlations (1985a:105):

(2) Diminishing linguistic distance between X and Y
a. X # A # B # Y
b. X # A # Y

c. X + A # Y

d. X # Y (analysis)
e. X + Y (agglutination)
f. Z (synthesis)

Essentially, this scale seeks to represent the significance
of the intervention of various linguistic units between any
two meaningful units X and Y. The intervening units may be
morpheme or word boundaries (+ and #, respectively), or other

morphemes or morpheme clusters (A or B). If more units, or

21

units of greater significance (compare d and e) intervene
between X and Y, the linguistic distance between them is then
greater by definition, according to scale (2).

Haiman demonstrates correlations between decreases along
this scale (in the direction from a to f) and decreasing
conceptual distance in several areas. Of these, we will now
briefly examine three: causation, transitivity, and
possession.

Causation is often expressed in more than one way in a
given language, and where forms differ in linguistic distance
along the scale (2a-f), conceptual distance varies
correspondingly. In Amharic, for example (after Hetzron

(1976)), we find:

(3) a. Abbat lagun saga AS—balla
father boy meat CAUS-eat
"The father forced the boy to eat the meat."
b. Abbat lagun s3ga A—balla
father boy meat CAUS-eat

"The father fed the boy the meat."

Here, we see a phonologically reduced form used to indicate
direct causation (3b), and the corresponding full form used
for indirect causation (3a). Since indirect causation

implies greater conceptual separation between the causative

action and the event caused (essentially making them two

22

separate events), we would in fact expect the phonologically
larger form to be correspondingly employed.

This is similar to the case in English, where (4a)
generally allows for greater temporal separation between

Herb's action and Les's death than does (4b):

(4) a. Herb caused Les to die.

b. Herb killed Les.

Example (4a) also allows for a wider range of scenarios than
does (4b); Herb may have hired a hit man or turned Les (an
escaped death row inmate) in to the police, rather than
physically doing Les in himself. Haiman cites supporting
data from several other languages, including Japanese and
Korean (1983:784-86).

In many languages, it is possible to mark a semantic
patient either with a direct case marker (accusative or
absolutive, for example), or with an oblique one. In such
cases, Haiman notes that the employment of an oblique case
may allow for some type of intermediate agency, while a
direct case usually does not. For example, we find in

Hungarian (after Comrie (1980)):

(5) a. Kohég-tet-em a gyerek—kel

cough—CAUS-lsg the child —COMMIT

 

23

b. KOhOg-tet-em a gyerek-et

cough-CAUS-lsg the child —ACC

Haiman notes that "both mean 'I make the child cough;' but
the first suggests that I do so by asking him to cough, while
the second suggests that I do so more directly, perhaps by
hitting him on the back" (1983:792). A similar situation

exists in French (taken from Hyman and Zimmer (1976:193)):

(6) a. Je lui ai fait preparer la mayonnaise.

"I had him prepare the mayonnaise."

b. Je l'ai fait preparer la mayonnaise.

"I made him prepare the mayonnaise."

Again, employment of a direct case —- accusative in (6b)
-— implies a more immediate effect on the semantic patient
than does employment of an oblique case -— dative in (6a).
Furthermore, both (5) and (6) demonstrate further support for
the linguistic distance hypothesis; in both examples, the
phonologically smaller forms (5b and 6b) correspond to
conceptual closeness.

With regard to possession, Greenberg has suggested (as
quoted by Haiman (1983:793)): "in no language will the
linguistic distance between X and Y be greater in signaling

inalienable possession, in expressions like 'X's Y,‘ than it

 

24

is in signaling alienable possession." This, again, reflects
the idea that items which are more closely associated
(inalienable) will not be linguistically farther apart than
are items which are less closely associated (alienable).
Johnston (1981:217) provides relevant data from the

Austronesian language Nakanai:

(7) a. luma taku
house my
b. lima—gu

hand—my

Other languages which offer such contrasts include Chiricahua

Apache and er11e (Haiman 1983:794).

2.4 MARKEDNESS

Another area in which iconic principles are claimed to
operate in morphosyntactic structures is that of markedness.
Haiman (1980:528) notes that "categories that are marked
morphologically and syntactically are also marked
semantically." Presumably, there would be no reason to
assign morphosyntactic complexity if it did not somehow
mirror complexity at another level.

Generally speaking, for example, the positive,
comparative and superlative degrees of adjectives in a given

language show some increase in the number of phonemes they

25

contain: English big, bigger, biggest; corresponding

Colloquial Japanese forms ookii, moto ookii, ichiban ookii;

 

corresponding Tamil forms periya, mika periya, elEthilum

 

periya.

Greenberg has also pointed out (as quoted in Haiman
(1980:528)) that "there is no language in which the plural
does not have some non—zero allomorphs, whereas there are
languages in which the singular is expressed only by zero."
Plurality then, is a marked category, and thus, one tending
to involve greater morphosyntactic complexity.

We might note a potential danger of circularity in such
arguments as these; it would certainly strengthen this case
if we could somehow define "semantic complexity" before
correlating it with "morphological complexity." As it
stands, semantic complexity remains a rather impressionistic
designation, although one which Haiman suggests has been
"universally assumed" (1980:528). He cites Greenberg (see

above) and Jakobson (1966) as examples of this assumption.

Nonetheless, some sort of definition —— perhaps in terms of
semantic distinctive features -- should be expected for this
term.

2.5 THE SPEECH OF CHILDREN
Slobin (1985) presents data from the speech of children
in various linguistic communities. He claims that children's

placement of certain operators within syntactic constructions

26

indicates iconic modification of parental language. In
particular, children seem to iconically model the scope of an
operator by their placement of it.

In the case of negation, for example, the placement of
negative markers is often modified, such that they appear
outside of the main clause. Slobin believes "children
indicate in their restructuring of parental languages that
the scope of negation should be the proposition, as indicated
by the verb as a whole, rather than any particular nonverbal
lexical item within the clause" (p. 222). He cites the case,
reported in Smoczyfiska (1985) of a Polish child who exhibited
"early sentence-external negation (e.g. Nie Basia §pi 'not
Basia sleeps' for Basia nie Spi)." This was followed by a
stage in which negation and verb were kept together, but
placed in sentence-initial position: Nie Spi Basia. Other
children are presumed to be indicating the same propositional

negation when they exhibit sentence-final negative placement

 

(e.g. mamusia kopac bedzie nie 'mommy dig will not'). Slobin
cites similar examples for children acquiring English,
Turkish, Japanese, French and Hungarian (pp. 223-24). He
points out that children's treatment of negation as a
propositional phenomenon makes them reluctant to permit its
presence to affect form or placement of other sentence
elements. This is demonstrated by, for example, the relative
slowness of children to acquire the obligatory genitive case

marking of the direct object of a negated verb in Russian;

27

children tend to put such objects in the accusative. Again,
data from several languages are cited to support the claim
(pp. 224-26).

The marking of conditionality is another area in which
Slobin suggests that iconic principles operate during
acquisition. The only available data are from Hungarian, in
which children tend to place the conditional morpheme outside
the verbal person/number affix, substituting beszélek volna
"speak+1sg COND" for beszélnék "speak+COND+lsg" (MacWhinney
(1973)). Slobin believes that this, again, is an
interpretation of a particular element (in this case, the
conditional marker) as a propositional-level element, and an
iconic indication of this interpretation in the element's
placement.

In general, Slobin suggests that acquisition data from
various languages indicate that children iconically mark on
both nouns and verbs the categories which they deem most
relevant to the information being communicated. The less
relevant any particular bit of information is considered to
be, the later it will be acquired (c.f. Bybee (1985). For
example, children tend to acquire verbal tense and person
affixes relatively early, but do not quickly adopt
stem-changes associated with either. Where stem—changes are
used to indicate aspect, however, children do "readily
acquire alternate verb forms" (Slobin 1985:231) -- even

overgeneralizing at times to use stem changes where

 

28

affixation is normal in the adult system. If the fact that
children pay greater attention to morphological markings for
one category than for others indicates that they consider
that category more immediately relevant, then their
employment of markings for that category may be termed
iconic.

With regard to sentential valence, Slobin notes that
children tend to give early attention to actual nominal
participants represented in a proposition, rather than to any
verbal morphology which might be used to indicate valence.
This indicates, he concludes, that children treat "the entire
configuration of noun and verb as the domain of valence,
rather than the verb itself" (pp. 232—33). This seems to be
a rather obvious example of diagrammatic iconicity as
discussed in 1.2.2, above, and it is certainly not limited to
the speech of children. The important consideration here,
however, is that children manipulate this fairly obviously
iconic strategy before they give attention to the less
isomorphic verbal morphology. Iconic representation, then,
is apparently more salient to the developing language-user.

Another interesting topic which Slobin addresses is a
principle of analyticity, which children may make use of to
express complicated notions by combinations of unitary forms,
rather than by complex forms. He cites the tendency of
children acquiring Romance languages to prefer analytic

PREPOSITION+PRONOUN forms over those which conflate

29

possession with person, number and gender (e.g. in French the
analytic de moi "of me" over the conflating forms mon/ma/mes
"my"). Such usage moves towards a one—to-one correspondence
of form and meaning (isomorphism), and is therefore deemed
more iconic.

For Slobin, the importance of all these data is that
they seem to indicate a tendency of children towards "making
their language temporarily more iconic than that of the
speech community" (p. 229). He interprets this tendency as
evidence that "iconic principles have a deeply-rooted
ontogenetic basis" (p. 229). Of course, if he is correct in
asserting that children naturally tend to make their
languages more iconic, then his data provide support for
claims such as those of Payne (1988) regarding iconic
principles and general cognitive human functions. The
apparent tendency of adult languages to disregard these
influences to any extent may then be seen as a result of
traditionally recognized diachronic processes such as sound
change and analogic change, which probably should be counted
among forces which accidentally oppose iconic pressures in
linguistic systems. Since such diachronic influences are
prevalent, fairly wide divergence of adult languages from

iconic coding relationships should not be overly surprising.

2.6 DEVELOPMENTS OVER DIACHRONY

In a paper that is "basically oriented towards

30

diachrony," (p. 278), Greenberg (1985) examines iconic
phenomena arising from metaphorical extension of deictics.
In particular, he is concerned with extension which involves
the use of originally spatial deictics to indicate temporal
relations, as well.

For example, a distant demonstrative may be
metaphorically extended to refer to past time. The issue, it
seems, is that both categories refer to that which is distant
(far from the speaker), and often invisible, as well.

Further support for this idea is provided by the fact that a
third person pronoun or article may also be derived from a
distant demonstrative; again, this is a category of entities
normally thought of as far away or absent, from the point of
view of a speaker.

In addition to the use of a far demonstrative to
indicate past, many languages employ a near demonstrative to
indicate future events. Thus, Old Irish ee, the near
demonstrative, refers to what follows, while elp, the far
demonstrative, refers to what precedes. Although it would be
logically possible, and explanations of metaphorical
motivation could certainly be constructed, Greenberg
nonetheless knows of no system in which a contrary rule
operates (i.e. use of far demonstrative for future, near for
past) (p. 285).

A related phenomenon is the use of far and near

demonstratives to indicate the equivalents of the English

 

31

"the former" and "the latter," respectively. In languages
employing such systems, the nearer demonstrative always
refers to the last—mentioned item, so that one gets the
impression of moving backwards in time from the moment of
speaking. This phenomenon occurs in German dieser/jener
"this/that," as well as in French, Hindi, and numerous other
languages (Greenberg 1985:285-86).

We might note that the iconicity involved in these cases
is of the system-internal sort described by Giv6n (1985), in
which the prior usage of a particular linguistic form within
its own linguistic system is the characteristic which enables
that form to iconically represent other referents. Thus, the
only reason that iEEEE may be iconically extended to serve as
both far demonstrative and indicator of past statements is
that, within the Modern Standard German system, speakers
agree on that particular phonological form as the far
demonstrative to begin with. However, the fact that both
functions are performed by a single form is iconic,
independent of the particular system in which this occurs.
Viewed in this light, the phenomenon is another instance of
projection of a cognitive principle of iconicity into the
linguistic system.

There is a danger here, however, of so extending the
term "iconicity" as to make it meaningless. Technical
vocabulary is useful up to the point that it serves to

differentiate; if a particular term comes to refer to

 

 

32

everything, it is no longer precise enough to be helpful. On
this basis, we might object to Greenberg's broadening of
"iconicity" to include metaphoric extensions.

On the other hand, Greenberg certainly has a point in
noting the essential similarity of iconicity and metaphor -—
the two phenomena do seem to involve similar representational
principles. Thus, however we eventually choose to label
them, these apparently related features of human language
together point to the existence of underlying cognitive

strategies which enable the operation of both.

2.7 PROCESSING STRATEGIES

In his (1987) article "Processing Strategies: A
Psycholinguistic Neofunctionalism?" Gary Prideaux endeavors
to show similarities between some cognitive strategies now
posited by psycholinguists and earlier concepts advanced by
the Prague School. Although he does not employ the term
"iconicity," he is describing syntactic indications of
conceptual strategies. Thus, his examples are of interest to
us in this context. The three strategies which he deals with
are labelled "Given-New, Closure, and Bracketing" (1987:300).

The Given—New strategy dictates that information which
is assumed commonly available to speaker and listener in a
particular context (Given) will be "systematically separated
from New information (that known only to the speaker)" (p.

301). Furthermore, this strategy calls for Given information

33

to precede New, unless "special grammatical devices" are used

to indicate that the expected pattern is not being conformed

to. For example, Prideaux offers the sentence:

(8) I sent the flowers to Sue.

as an appropriate response to:

(9) Who did you send the flowers to?

but not as an appropriate response to:

(10) What did you send to Sue?

Example (9) treats "the flowers" as Given information,

whereas (10) does not. Therefore, Prideaux claims that (8)

can only be an appropriate response to (10) if it is

specially marked; in this case, PHONOLOGICAL marking would

result in:

(11) I sent the FLOWERS to Sue.

which would indicate that the normal Given—New strategy was

being violated.

Prideaux formulates the Closure strategy as follows:

"In processing a particular unit (phrase, clause, etc.) the

 

34

hearer attempts to gain closure on that unit at the earliest
possible point" (p. 303). As an example, he notes that
hearers must stop mid—stream and reanalyze when presented
with a sentence like: "The guest expected to be absent
arrived." The reason for this, he claims, is that hearers
tend to project closure on the sentence before its end
arrives. This conclusion is supported by various empirical
studies, which Prideaux alludes to but does not explicitly
cite.

The third processing strategy presented is that of
Bracketing, which states that "the hearer expects that when a
new unit for processing is encountered, it will be marked as
such" (p. 305). Prideaux points out the obligatory use of

the complementizer that in (12):

(12) That Sally has never eaten sushi is obvious.

and its optionality in (13):

(13) It is obvious (that) Sally has never eaten sushi.

The reason for the obligatory use of that in (12) is that the

bearer will not expect an initial subordinate clause, and

thus, the speaker will need to alert him of its presence.

If the ideas which Prideaux has presented are found to

stand up cross—linguistically, (which, of course, can only be

 

35

demonstrated with further data), these principles may in fact
prove to be useful exemplifications of how syntax may

iconically reflect cognitive processing.

2.8 THE LEXICON

In the final section (pp. 535-37) of his 1980 paper,
Haiman addresses the question of the relationship between the
relative size of a language's lexicon and the degree of
discoverable grammatical iconicity. Not only does a
(relatively) lexically—restricted language generally require ,
longer messages than does a more lexically—rich language, in
order to express the same idea, but the former is likely to
exhibit more iconic syntactic motivation, as well.

As an example, Haiman describes the situation which
exists between the social dialects of Dyirbal: Dyalnuy and
Guwal (data taken from Dixon (1971)). Dyalquy is a taboo
language, utilized in different social contexts than Guwal;
Dyalguy has both a smaller inventory of lexical items and a
higher incidence of iconic motivation. Whereas, for example,
categories of repetition and distributivity may be indicated
by suppletive forms in Guwal (EEBEE "look," gunin5u "search")
these same categories are often indicated in Dyalguy by
morphological devices such as reduplication (bunman "ask to
accompany oneself," bunma+bunman "keep asking to accompany
oneself") or aspectual suffixation (n9uRiman "look,"

n9uRima+rind9an5u "search").

 

36

Similarly, English translations of Eskimo words for
"snow" such as "packed snow" or "drifting snow" demonstrate
the same sort of phenomenon: "redressing a deficiency in the
lexicon by a greater perspicuity in compounding" (Haiman, p.
537). Haiman suggests that one might also expect to find
greater iconicity in pidgins, compared to the languages whose

vocabularies they have "radically simplified" (p. 537).

2 . 9 DISCUSSION

The preceding sections have shown that various analysts
have attempted to demonstrate the presence of iconic
motivations in a variety of areas of morphological,
syntactic, and discursive coding in natural languages. Some
of these data seem, admittedly, less convincing than others.
However, it is hoped that their combined weight —— if not
each argument individually -- will serve as convincing
evidence that the theorist must admit the possibility of

iconic motivations for various linguistic constructions.

3. AN ANALYSIS OF ICONICITY IN WORDSWORTH'S POETIC SYNTAX
3.0 INTRODUCTION

William Wordsworth is numbered among the greatest of the
English poets. His influence is such that it has been
claimed "the history of the poetry of the last two hundred
years seemed most easily comprehensible when considered as a
whole and interpreted as beginning with Wordsworth" (Rehder,
1981:15). This poet attempted to achieve something new with
his work, and his success has had a remarkable impact on
poetic art.

In order to understand some of the syntactic effects
which we will examine in this Chapter, we must first be aware
of what Wordsworth was attempting to do with his verse. This
Section will provide a brief introduction to the poet's
goals.

Much of Wordsworth's artistic inspiration derived from
Milton. In particular, Wordsworth was impressed with
Milton's expressed desire "to be an interpreter and relator
of the best and sagest things among mine own citizens
throughout this island in the mother dialect" (Greenbie,
1977:17-18). As we shall see in Section 3.3, below,
Wordsworth disagreed with his predecessor as to the nature of
"the best and sagest things," but his desire to handle such
"things," whatever they would turn out to be, certainly
places him in Milton's company with respect to this common
goal.

Of particular significance for the present study,

however, is Wordsworth's desire to go beyond Milton in his

37

 

38

presentation of ideas "in the mother dialect." Where Milton
had utilized an educated, stylistic language, Wordsworth
desired "to choose incidents and situations from common life,
and to describe them... in a selection of language really
used by men," and, later, "to adopt the very language of men"
(Wordsworth, 1896:Vol. V:192, 198).

The effects which these efforts achieved have been
described as Wordsworth's "plain, pluralizing modez"
"plain," because the poetry lacks the traditional trappings
of verse, including both archaic syntax and flowery
vocabulary; "pluralizing," because the effect of this
plainness is to make Wordsworth's art accessible to all who
care to read it. In this Chapter, we will explore some
iconic features of Wordsworth's syntax which help to explain

these effects.

3 . 1 NOMINALS

Wordsworth's poetry is firmly grounded in the experience
of the common man, and one important area in which this can
be observed is that of his nominal constructions. The poet
does not seek to create a world for his reader's imagination
-— as does so much of literature. Rather, he attempts to
help his reader to discover an already existing world -- the
world of the familiar, the common, the world of everyday
life. His goal in referring to various participants and

props in the drama he lays out is that we should both

 

39

recognize them and incorporate them quickly into the
experiences in which they play their roles; descriptions of
the players and their surroundings are incidental, rather
than central.

Thus, we find a proliferation of nominal words, without
any accompanying proliferation of adjectival modifiers. The
nouns themselves are familiar to all, and it is the
experience of many common objects together, rather than the
specific characteristics of any individual object, that the
poet is concerned with. Ward (1984:55ff) deals extensively
with what he calls Wordsworth's "plain nominals,"
characterizing them as "plain language" (p. 57) and
"conventional, ordinary words" (p. 58).

In this section we will investigate some syntactic
characteristics of these "plain nominals," and will suggest
that iconic aspects of these characteristics may be found.

Giv6n (1987) describes a scale for correlating the
predictability of reference to a particular participant in
discourse with the type of nominal used to present it in a
given instance. The claim which Giv6n makes is that a more
predictable topic of discussion (that is, one which is more
central to the discussion) often requires less description to
be understood than does a topic which is unexpected or
non-central to the discourse; that which is unexpected
requires extra clarification. This, then, is an iconic

relationship between syntactic coding and participants being

40

coded. The following scale (adapted from Givdn's work) is
prOposed to demonstrate the correlation between the
predictability of a t0pic and the type of nominal marker
likely to be used for its presentation:
(14) TOpic Predictability and Phonological Size

MOST PREDICTABLE TOPIC

a. zero anaphora (Omission)

b. pronoun

c. definite noun phrase

d. indefinite noun phrase

LEAST PREDICTABLE TOPIC
This scale is intended to express a universal hierarchy of
coding devices, valid for any language which makes relevant
coding distinctions.

In order to check the viability of these generalizations

for a given text, we need a way to describe how predictable a
given topic is in that text; Giv6n proposes a referential
distance test, which assumes that tOpics central to a
discussion are simply mentioned more often than others. To
apply the test, one simply counts the number of clauses
between references to particular participants, and then
checks for correlations between number of clauses elapsed
since the last mention of a participant and the type of
device used to code that participant in that instance. An
upper limit on counting of 20 clauses is arbitrarily

assigned, largely to keep numbers managable. Giv6n points

 

41

out, however, that referential distance measurements for
nominals characteristically employed to "return a topic into
the register after a relatively long gap of absence"
generally yield values of 15-17 clauses, and that "this value
is already biased upwards by the arbitrarily—assigned 20
clause value" (1983:36:fn#6). Thus, the 20 clause limit may
actually be too high, rather than too low.

Scale (14), presented above, is generally shown to be
accurate, in that coding devices which are lower down on the
scale tend to correlate with greater referential distances
than do the higher devices. Giv6n (1983), for example, is a
collection of papers presenting studies of these correlations
for Japanese, Amharic, Ute, Biblical Hebrew, Spanish, written
and spoken English, Hausa and Chamorro.

The referential distance test is primarily intended to
be applied in narrative discourse. Thus, we might expect to
encounter difficulties in finding consistent correlations if
we apply it to Wordsworth's poetry, which is generally fairly
meditative and introspective, and not very narrative in form.
For the current study, however, the test has been applied to
two different sections of Wordsworth's poetry, and the scale
has been found to be generally accurate in both situations.
The more predictable a given topic, the less coding material
is needed to represent it in a given instance, apparently
independent of how narrative the passage is in form.

However, the results indicate that more than simple topic

 

. Va". (1 y 390

    

;lrn1m03

. .
l-
a
n
-_ l
-. f| '_-
.
\
.' .r
. "
I _ _. .
_ J

42

predictability is at work. The remainder of this section
presents findings for a relatively introspective and a
relatively narrative passage, both from Wordsworth's work.
Table I presents referential distance values for a fairly
meditative section of Wordsworth's work, where we might
intuitively expect to find a large amount of interference
from artistic freedom of expression. The section from which
these data were taken is Book I, lines 340-56 of the
introspective autobiography "The Prelude Or The Growth of a
Poet's Mind." (Information about all poems cited in this

thesis is provided in the Appendix.)

TABLE I
Referential Distances in a Meditative Section

of Wordsworth's Work

 

Coding Type No. Occurring Ave. Ref. Dist.
a. zero anaphora 2 1.0
b. all pronouns 10 2.1
c. Def. NP 15 18.9
d. Indef. NP 14 18.6

These results do, in fact, generally correlate with
those found by Brown (1983) for written English narrative.
Brown's results are partially summarized in Table II for

comparison with those presented above:

 

 

43

TABLE I I
Referential Distances for Written English Narrative

from Brown (1983)

 

Coding Type Ave. Ref. Dist.
a. zero anaphora 1.00
b. unstressed pronouns 1.72
c. demonstratives alone 2.27
d. Def. Art. + NP 16.66
e. Indef. referential 19.17
f. generics 19.23

Brown's data and these data both show a marked increase
in average referential distance between the larger devices
(definite and indefinite NP's) and the smaller devices
(pronouns and zero anaphora). It seems that these coding
types naturally fall into two distinct categories, which may
be called "full" and "reduced" nominals.

If we rework the data of Table I on the basis of just
these two categories, we obtain the averages presented in

Table III:

 

44

TABLE III
Referential Distances for Full and Reduced Nominals

from Table II

 

Coding Type No. Occurring Ave. Ref. Dist.
a. Reduced NP 12 1.8
b. Full NP 29 18.8

A particularly interesting fact about these data is that
the average referential distance for all full nominals so
closely approaches the upper limit of 20 which was allowed in
the original counting. This suggests that full nominals,
whether definite or indefinite, are most typically used to
introduce completely new topics -- at least in this
meditative passage -- and that further reference to any
topic, if further reference is made at all, will usually be
assigned to a reduced nominal coding device.

The fact is that Wordsworth's most meditative verse is
full of nominals which have Virtually no topicality at the
points where they are introduced; he is constantly
introducing new background and participants, only to mention
them once and let them fall from sight again immediately.
These nominals refer to things which are known to every
person -- things which are common to everyone's experience.
They are seldom greatly elaborated —— adjectives are present,

but not abundant -- because they are things which the poet

45

simply wants to call to mind, but not to describe. Here,
then, we see quantitative evidence of some iconic syntactic
features central to Wordsworth's "plain, pluralizing mode"
(Ward 1984:63).

As we have already noted, the hierarchy of nominal
coding devices proposed by Giv6n is intended primarily to
apply to narrative discourse. The scale is expected to yield
better correlations when participants are being introduced
and dealt with in some logical flow of description. We have
seen, however, that some measure of correspondence still
exists, even when referential distances are compared within a
rather "artistic" piece which hardly qualifies as narrative.

Some of Wordsworth's poetry, however, does take a more
narrative approach to its description, albeit with some
editorializing and a high degree of artistic stylization on
the part of the author. One such poem is "The Idiot Boy," in
which Wordsworth describes a mentally-impaired child's
pony-ride to town to get a doctor late one night. Table IV
presents referential distance figures for nominal coding

devices in lines 1-100 of this poem.

 

46

TABLE IV

Referential Distances in Wordsworth's Poetic Narrative

 

Coding Type No. Occurring Ave. Ref. Dist.
a. zero anaphora 20 2.0
b. pronoun (all types) 48 3.1
c. proper name 24 3.4
d. Def. NP 65 15.4
c. Indef. NP 54 16.9

Here again we see a fundamental confirmation of the
scale of correlations proposed by Giv6n. Smaller coding
devices iconically indicate smaller intervals between
references to a given participant, while larger devices
signal that a participant is more unexpected, or less
topical, in the discourse.

It is interesting to note that proper names, included as
a separate category in Table IV because of their frequency in
"The Idiot Boy," seem to fall much closer to pronouns than
to other definite noun phrases. Perhaps various constraints
such as metrics and rhyme are working together here to cause
Wordsworth to treat proper names as simply alternate
pronouns; the poet can choose either pe or Johnny to refer to
the idiot boy himself, and of course the former has only one
syllable, while the latter has two. Thus, metrical

considerations may have a strong influence here.

 

47

Additionally, Wordsworth often reintroduces a character by
proper name at the first mention of that character within a
stanza -- almost as if each stanza began a new story. Since
each stanza is only five lines long, this produces a rather
large number of proper name uses, within a relatively small
area. Hence, there is a small referential distance for this
category.

Table IV also reveals somewhat lower average referential
distances for full nominals than were found in Table I. This
is probably attributable to the fact that, in the more
narrative passage, Wordsworth is concerned with following a
plot line, rather than simply guiding the reader's thoughts
to and fro. Thus, the participants are simply more
continuous in the narrative verse, such that all types of
nominals are relatively more likely to code predictable
topics.

Of course, adopting such an argument leaves us with the
question of why pronouns show an increase in referential
distance in Table IV, if we predict a general decrease for
narrative passages. Two possible explanations might be
suggested: first, "The Idiot Boy" adheres to rather rigid
metrical and rhyming patterns, which certainly have an impact
on syntactic arrangements, as we noted in Section 1.5;
additionally, the very small number of participants in the
poem (essentially 3) makes it possible for pronouns to refer

back over large distances without fear of creating confusion,

 

 

48

whereas the absence of central participants among the
plethora of nominals in the more meditative passage first
examined creates the need for pronouns to be utilized only in
close proximity to their antecedents for the sake of clarity.

In general, we find that Wordsworth's nominals do seem
to exhibit the sort of correlations predicted by Giv6n, in
passages of both narrative and introspective character.
Participants which are highly topical at any given point in
the discourse are generally coded with small nominal coding
devices, while less topical participants require larger, more
explicit coding devices.

We have already observed that one of Wordsworth's goals
in writing is to make his work easily accessible to the
common man, and that he utilizes abundant references to
commonplace things as one strategy for achieving this effect.
We have suggested that such references involve background and
participant nominals which have virtually no topicality at
the points where they are introduced, and which thus tend to
be coded by full nominals.

There is an apparent contradiction here. Such nominals
are at once highly familiar to the reader and lacking in
topicality in the discourse being presented. It seems that
Wordsworth deliberately manipulates these two phenomena in a
way that brings their theoretical conflict into a subjective
harmony for the reader.

Ward (1984:58) has suggested that Wordsworth's verse

 

49

tends to contain many items, "of which none is central."

This impression is apparently created by the juxtaposition of

reduced nominals —— nominals which are high in topicality in
the discourse at hand —— with full nominals -- nominals which
are high in a different sort of topicality —— that of human

experience. In his work, the poet makes them seem equivalent
in importance, although their topicality derives from

different sources.

3.2 PREVERBAL OBJECTS

In Wordsworth's work, one finds a number of instances in
which the direct object of a transitive verb occurs in
preverbal position. This is neither normal in English syntax
nor particularly common in Wordsworth's work, so that the
reader might expect to encounter some initial difficulty in
recognizing the grammatical relations of preverbal objects
within the sentences which contain them. Since preverbal
position in English is normally reserved for grammatical
subjects, we would naturally expect that preverbally
positioned objects could be misperceived as subjects —— at
least on first reading.

This, however, is generally not the case. In this
section, we will examine several instances of preverbal
occurrences of direct objects, and suggest that specific
features of the syntactic constructions involved serve to

indicate iconically the status of the preverbal nouns in

or!

Hb '
as):

 

 

50

question as unambiguously objects.

It should initially be noted that English syntax (or at
least, poetic license) allows for the placement of nominals
with several types of grammatical relations in either

preverbal or postverbal postion. Thus:

(15) To both I listened

(Prelude 1:57)

(16) While upon the fancied scene I gazed

(Prelude 1:76)

(17) In the castle he's pursuing

(Idiot Boy 229)

(18) ...and in a faltering voice,
Whose tone bespake reviving interests
Till then unfelt, he thanked me

(Prelude IV:463—65)

In examples (15)-(18), the preverbal nominals of
interest occur within prepositional phrases. 29 in (15) and
epee in (16) are particles which introduce phrases that
co-occur with the verbs listen and geee, respectively.

Examples (17) and (18) demonstrate two different functions of

 

51

the prepostion lp: to introduce a locative expression in the
former and an adverbial of manner in the latter.

Of course, it is perfectly permissible in English
grammar for these categories to be expressed preverbally; the
point here is simply that their grammatical relations are
clearly identified by the prepositional markers te, welle,
and le.

English direct objects normally correlate with semantic
patients, and of course, a means is provided by English
grammar for patients to occur preverbally, as well. The
typical vehicle for this is the passive construction, which

is also found in Wordsworth, as in examples (19)-(21):

(19) The trees (her first—born child being then a babe)
Were planted by her husband and herself

(Grasmere 391-92)

(20) ...I feel
That an internal brightness is vouchsafed

(Grasmere 674-75)

(21) ...till choice was made
Of a known Vale

(Prelude I:71-72)

'.:.'..l'_‘(i

    

earl:

 

52

In these and all examples of English passive
constructions, unique coding of the verbs and the optional
py+agent constituent serve as iconic indications of
unexpected semantic information.

Having dealt with preverbal occurrences of other
nominals, we may now examine those nominals which occur
preverbally, signify semantic patients, and are pet coded as
the subjects of passive constructions. One might expect that
a reader could experience difficulty in recognizing the
grammatical relations of such nominals, in a language fairly
free of case markings, as is English. However, we will find
several iconic indicators helpful in identifying the
syntactic relations of these nominals.

Dillon (1978) suggests that, in reading literature, one
may employ "four kinds of strategies for identifying
[preverbal] Subjects and Objects: strategies based (1) on
serial order, (2) on pronominal clues, (3) on semantic
compatibility (enriched by context), and (4) on perception of
theme" (p. 22, enumeration added). The third and fourth
types are actually the most important, but they do not fall
within the realm of what we are investigating here. However,
we will see that syntactic clues contribute to our ability to
use strategies of types 1 and 2 in correctly identifying the
grammatical relations of preverbal objects.

Dillon's type 1 strategy applies to cases in which both

subject and object in a sentence are preverbal, yielding a

53

construction of the form:

(22) NP - NP - V.

In such cases, strategy 1 (Dillon 1978:9) asks us to assume
that the object will be expressed first, and the subject

second:

(23) NP(O) - NP(S) - V.

This type of construction is not common in Wordsworth, but
where an NP — NP string does occur preverbally, Dillon's rule

generally seems to hold:
(24) When this Vale
We entered

(Grasmere 170—71)

(25) For peace they have

(Grasmere 282)

More often, however, we find that Wordsworth employs a

construction of the form:

(26) NP - Aux - NP - V.

54

In these situations, we also find a serial order clue, which
is that once again the grammatical direct object will be the
first NP, and the subject will be the second. Such

constructions occur in examples (27) and (28):

(27) Friends shall I have at dawn

(Grasmere 515)

(28) The owl that gives the name to Owlet-Crag
Have I heard whooping

(Grasmere 521—22)

The occurrence of either of the syntactic patterns
discussed above serves as a fairly consistent indicator of
preverbal object positioning. Thus, just as passive
constructions iconically indicate semantic patient status of
grammatical objects, so are these constructions employed by
Wordsworth as iconic indicators of semantic patient status of
preverbal nominals.

It should be noted that iconicity of the type being
discussed here is not the sort in which syntactic patterns
actually mirror semantic information, as was the case with
the noun phrase phenomena described above. Rather, syntactic
patterns are serving as indicators that the sentences in
question are presenting information differently than what is

normal and expected. No inherent quality of patterns (23)

 

55

and (26) helps us to recognize the grammatical relations of
their parts when we encounter them, but their consistent use
within Wordsworth's work when objects occur preverbally makes
them iconic indicators, within Wordsworth's system, of the
semantic information we have been discussing.

Dillon's second type of strategy involves the
case—marking of pronouns. Quite simply, an author may
identify which noun phrase in a construction is the object by
choosing a coding device for either subject or object or both
which will receive morphological marking to identify its
grammatical role unambiguously. This may be true regardless
of whether the grammatical subject of a sentence also occurs
preverbally —- as in (22) —— or not.

Actually, this strategy is relevant in virtually all
examples of preverbal objects in Wordsworth's work; in
addition to the examples given above, example (29)
illustrates a case-marked preverbal object, while (30) and

(31) illustrate case-marked subjects in different positions:

(29) but me hath Nature tamed

(Grasmere 726)

(30) cheap matter offered they

(Prelude I:529)

56

(31) No doubt too he the moon had seen.

(Idiot Boy 444)

This effect may be made even stronger by using a reflexive
pronoun, which, unless it serves as an intensifier, will of

necessity will be the object of the sentence anyway:

(32) While Earth herself is adorning

(Immortality 43)

Of course, case-markings on a subject, as in (30) and
(31), do not immediately identify the object; they leave the
reader to infer its role from the fact that the subject slot
in the sentence is clearly filled by the overtly marked
participant. But this is not much of an intuitive leap, and
it seems doubtful that it will often be missed. These clues,
then, may also be seen as iconic syntactic indicators of the
grammatical relations of preverbal objects.

Although semantic clues are most certainly primary in
our identification of preverbal objects, the syntactic and
morphological features which have been presented here can be
found to be of relevance in almost every occurrence of a
preverbal object in Wordsworth's work. Here, then, is
another area in which grammatical patterns correlate with
semantic information in what may be called an iconic

relationship.

57

3 . 3 SENTENCES

One of the characteristic features of Wordsworth's work
—- and one major reason that his syntax provides interesting
areas for linguistic study -- is that his sentences tend to
be extremely long. One sometimes gets the feeling that the
poet has, in fact, perhaps forgotten to utilize periods, or
that he quite possibly simply doesn't like them. Instead,
one finds multitudes of subordinate and modifying clauses,
along with generous use of punctuation to extend sentences;
colons, semi-colons and dashes abound.

The flow of Wordsworth's thought, however, is not
hampered by his seeming abhorrence of complete stops.

Rather, the punctuation seems to facilitate the flowing
together of thoughts which somehow ought to be joined —-
concepts which belong together, and which would suffer a loss
of intensity if broken up and separated into unitary
propositions.

This is all the design of the poet, of course. Not only
are such devices common in literature of Wordsworth's time,
but they seem to be employed by Wordsworth himself to achieve
particular effects, which we will seek to identify here.
Often, for example, Wordsworth seeks to link his thoughts
together syntactically to express their essential unity in
his own mind, and to allow the reader to pursue them in a
natural flowing of his own thought process; essentially,

Wordsworth is aiming to mimic the "stream of consciousness"

58

by means of his syntax.
Wordsworth alerts us to the fact that he will use his
verse as a tool to examine the mind. He writes:
(33) Urania, I shall need
Thy guidance, or a greater Muse, if such
Descend to earth or dwell in highest heaven!
For I must tread on shadowy ground, must sink
Deep - and, aloft ascending, breathe in worlds
To which the heaven on heavens is but a veil.
...we look
Into our Minds, into the Mind of Man —
My haunt, and the main region of my song.

(Excursion 25—30, 39—41)

Here, Wordsworth warns that he intends to attempt to fathom
"the Mind of Man," an attempt not yet made by poets in
general. He joins the ranks of those who essentially cast
away the reverences and fears of human history in turning
inward upon themselves, and upon the internal workings of
humanity.

Wordsworth's turn inward upon himself, then, is mirrored
in the manner in which he constructs his sentences; he
attempts to model the turning and twisting of his own
thoughts in the complex syntax with which he presents his
work. This may be seen, too, as another attempt to make his

work more accessible to the common man; if the syntax really

59

mirrors the workings of human consciousness in any way, then

his readers will certainly find it easy to process and fairly
accessible, at least in terms of the syntactic aspects of the
work.

Additionally, this may be seen as another area of
iconicity in Wordsworth's grammar. Here, the poet's
perception of the way the human mind deals with the world
around it is reproduced in both syntactic arrangements and
the grouping and linking of individual propositions into
larger units within the total discourse.

In this discussion, we will suggest some ways in which
Wordsworth's deliberately contrived syntactic patterns do
seem to serve as iconic indicators of the flow of human
consciousness.

Let us begin by looking at a fairly typical sentence
from Wordsworth's extended autobiographical poem. This
particular one comes from Book IX of the "Prelude," entitled

"Residence in France:"

(34) Where silent Zephyrs sported with the dust
Of the Bastille, I sate in the open sun,

And from the rubbish gathered up a stone,

And pocketed the relic, in the guise

Of an enthusiast; yet, in honest truth,

I looked for something that I could not find,
Affecting more emotion than I felt;

For 'tis most certain, that these various sights,

60

However potent their first shock, with me
Appeared to recompense the traveller's pains
Less than the painted Magdalene of Le Brun,
A beauty exquisitely wrought, with hair
Dishevelled, gleaming eyes, and rueful cheek
Pale and bedropped with everflowing tears.

(Prelude IX: 67—80)

This sentence is set off as a poetic stanza unto itself, and
is actually probably a little bit longer than the average
sentence in Wordsworth. However, there are many others as
long or longer; here, one sentence spans 14 poetic lines,
while its four immediate successors span 15, 11.5, 4, and
14.5 lines, respectively.

This sentence contains 14 clauses (i.e. main verbs,
counting two which are ellipsed, in lines 75 and 77) -— six
independent and nine dependent. Additionally, there are four
conjunctions —— "and ...and ...yet ...for" —- and one lengthy
appositive (the final three lines). At the outset, at least,
the sentence appears to be more or less narrative in
progression, and this gives it most of its logical structure.
As the poet continues, however, things become increasingly
contemplative, until the activity is gone entirely, and only
meditation remains.

In the process, the poet has strung together so many

clauses of such different focus that one would have rather

61

great difficulty explaining what the central proposition of
the sentence is supposed to be; the sheer bulk of what is
presented causes us to recognize none of the content as any
more critical than the rest, but all as somehow of equivalent
importance.

This is probably Wordsworth's intent. Just as we noted
with regard to noun phrases that there often seems to be no
central participant in Wordsworth's discourse, so we find
here that the sentence contains no central proposition.
Instead, a sentence consists of many propositions which
coordinate with, or depend upon, one another, but none of
which seems to command greatest attention. In this way, the
poet relativizes the importance of each proposition
expressed; none is essentially more important than the
others, and all seem to share equal prominence in the poet's
mind. Via his syntax, Wordsworth encourages his readers to
share in this sort of unitary treatment of experience and
thought.

We will now further examine Wordsworth's use of several
devices —— punctuation, subordination, apposition,
parallelism and ellipsis —— in order to see how these work
together toward the overall effect which we have been
describing.

3.3.1 PUNCTUATION
In order to get an initial feeling for how Wordsworth

achieves these kinds of effects with his sentences, let us

62

look at Table V, which presents data concerning the

punctuation used in the famous "Ode: Intimations of

Immortality from Recollections of Early Childhood:"

TABLE V

Punctuation Marks of the Immortality Ode

TYPE OF MARK

Comma

Semi-colon

Colon

Dash

Exclamation Point
Period

Question Mark

TOTAL

NO. OCCURRING

137
34
13

7
16

12

222

This ode (hereafter "Immortality") has a total of 203 lines.

From Table V, we find that sentence-ending punctuation marks

(periods, exclamation points and question marks) occur 31

times, so that the poem averages 6.5 lines per sentence.

(Actually, the average might be a bit higher, since the

exclamation point is sometimes used after brief interjections

which the poet doesn't seem to consider separate sentences,

since he follows them with lower-case letters.) The poem is

divided into ten sections of unequal length, so that each

 

63

section averages about 20 lines and 3 sentences.

Most importantly, we find that colons and semi—colons
together comprise 47 of 222 total punctuation marks. This
means that about 21% of Wordsworth's punctuation marks are of
the type which deliberately prolong a sentence at points
where it would almost always be possible to come to a
complete stop. The poet extends the boundaries of the
sentence to include multiple propositions.

3.3.2 SUBORDINATION

Example (34), above, is a good illustration of how
Wordsworth often makes extensive use of subordination within
a sentential unit. This allows him to present a great deal
of background information within the normal flow of thought.
Thus, we find adverbials of time, location, and manner, as
well as parenthetical comments about the author's feelings,
attitudes, and opinions.

Another example of this extensive presentation of
background information through adverbials is found in

"Michael: A Pastoral Poem:"

(35) And hence this tale, [while I was yet a Boy
Careless of books, yet having felt the power
Of Nature, by the gentle agency
Of natural objects,] led me on to feel

For passions [that were not my own] and think

64

(At random and imperfectly indeed)
On man, the heart of man, and human life.

(Michael 27-33)

The parentheses are Wordsworth's own, but the brackets
have been inserted here to highlight the adverbial digression
and the relative clause. This digression is of particular
interest, since the poet chooses to employ it in a position
where it separates the subject and verb of the main clause by
three full lines. It is as if a thought of tangential
importance struck him all at once, and he simply wrote it
down, so as to keep from forgetting it, and then continued
where he had left off.

And yet, that which seems syntactically to be background
or less vital information is clearly just as central to the
poet's message as is the main sentential matrix. If we were
to remove the material in Wordsworth's parentheses and in the
brackets, we would find that only 22 of the original 55 words
remained. Perhaps the unmodified main clause is best
considered a framework for supporting the poet's
meditiations; clearly, each needs the other to complete a
coherent presentation of the message.

3.3.3 APPOSITION

Another device which Wordsworth sometimes employs to

increase the amount of material which a sentence can hold is

apposition. Examples (36) and (37) demonstrate the use of

 

65

this device:

(36) ...this infant sensibility,
Great birthright of our being, was in me
Augmented and sustained.

(Prelude II: 270-72)

(37) Those fields, those hills...
...were to him
A pleasurable feeling of blind love,
The pleasure which there is in life itself.

(Michael 74—77)

Here, again, Wordsworth associates ideas by
interspersing related thoughts within the framework of his
main clauses. By the use of this device, the poet seems to
attempt to mirror the mental process of constantly shifting
his focus on what he is describing -- refocusing,
rebalancing, refining the mental image he seeks to produce,
all within the natural flow of thought.

3.3.4 PARALLELISM

At times, of course, all of Wordsworth's seeming
digressions may cause the reader to lose the train of thought
entirely. While this, too, may be a natural mental process
(and therefore, one which Wordsworth might seek to reproduce)

it is nonetheless not usually the poet's goal to confuse his

 

66

audience. Thus, he provides clues -- some syntactic -— to
help in making the proper associations.

One such clue is provided by syntactic parallelism, when
concepts which belong together are presented with similar
syntactic devices. This, of course, is an iconic use of

syntax to aid in semantic processing. For example:

(38) Well do I call to mind the very week
When I was first entrusted to the care
Of that sweet Valley; when its paths, its shores,
And brooks, were like a dream of novelty
To my half-infant thought; that very week,
While I was roving up and down alone,
Seeking I knew not what, I chanced across...

(Prelude V: 426—33)

Here, the poet has interjected yet another, four-line
adverbial, and perhaps fears that the reader will lose the
train of thought. In order to call attention back to the
original idea, he returns to a syntactic pattern from earlier
in the sentence: "the very week/ when I was first entrusted"
is recalled by "that very week/ while I was roving."

In this same example, Wordsworth also helps us to
identify background material by presenting it in consistent
forms. Two adverbials describing temporal setting both begin

with when, and two clauses describing the poet's activity

 

67

begin with the participles roving and seeking. Both pairings
help us to classify these modifiers together accurately, and
allow us to recognize more quickly that each presents
background for the main clause. Similarly, when we encounter
the verb chanced, we are likely to assume (correctly) that it
once again takes up the action of the central sentential
matrix, since it bears tense inflection matching that of the
rest of the primary narrative and does not appear to be
subordinated. All of these clues involve recognizing
morphological parallelism inherent in the structures which
Wordsworth has provided.

In this area of morphological and syntactic parallelism,
we see a typical example of the iconic principle "similar
form corresponds to similar function," which was discussed in
Section 2.2. The phrases "the very week" and "that very
week," however, go beyond the basic principle, in that they
are the result of deliberate choice on the part of the poet
to construct parallelism at a level higher than that of
obligatory morphology. Whether Wordsworth ever consciously
considered this principle or not, he seems to have made use
of its effects here.

3.3.5 ELLIPSIS

The syntactic devices which we have examined so far have
served two basic purposes -- the incorporating of multiple
concepts or propositions within a single sentential

framework, and the associating of those concepts with one

68

another. Now we will examine one more device -- ellipsis --

which Wordsworth sometimes employs to help perform the latter

of these two basic tasks.
Dillon (1978:123) presents
terms Wordsworth's "ellipsis of

main verbs." A typical case is

(39) [I] listen to the grave

Of dire enchantments faced

several examples of what he
semantically rather empty

example (39):

reports

and overcome

By the strong mind, and tales of warlike feats,

Where spear encountered spear, and sword with sword

Fought, as if conscious of

the blazonry

That the shield bore, so glorious was the strife;

Whence inspiration for a song that winds

Through everchanging scenes of votive quest

Wrongs to redress, harmonious tribute paid,

To patient courage and unblemished truth,

To firm devotion, zeal unquenchable,

And Christian meekness hallowing faithful loves.

(Prelude 1: 174-85)

Dillon suggests that we can best understand this passage

by supplying a missing main verb comes or came following

whence (1978:123). This does in fact make sense out of it

all, but the sentence is far from easy to parse, even then.

We could explain such ellipses with some principle

69

governing omission of semantically redundant material:
"information inferrable from context may be omitted." But
this would be foolish, since we have already spent great
effort showing that Wordsworth takes pains to include all
sorts of redundant and/or inferrable information. If
anything, he is normally overly explicit. So a few
relatively isolated omissions must be due to some other
consideration.

In fact, it seems likely that something related to
Haiman's linguistic and conceptual distance correlations is
taking place in such instances. By occasionally leaving out
sentential elements which can be inferred from context,
Wordsworth may be seeking to associate other parts of the
sentence closely by placing them close together. By omitting
the verb eeme in (39), Wordsworth places the "inspiration"
and its source in closer proximity -— and leaves it for the
reader to infer the exact relationship between the two. This
is greater conceptual closeness, iconically signalled by
proximity of the coding devices.

An even more interesting example of this phenomenon
comes from earlier in the same book of Wordsworth's

"Prelude:"

(40) But from this awful burthen I full soon
Take refuge and beguile myself with trust

That mellower years will bring a riper mind

70

And clearer insight. Thus my days are past
In contradiction; with no skill to part
Vague longing, haply bred by want of power,
From paramount impulse not to be withstood,
A timorous capacity from prudence,

From circumspection, infinite delay.

(Prelude I: 234-42)

The syntax of (40) is far from transparent, and it takes
some thought to work out the relationships in the second
sentence. Dillon suggests that we read it "pett a from b, c
from d, from e, f" (1978:123). This works nicely, with the
following correspondences: a="vague longing, haply bred by
want of power;" b="paramount impulse not to be withstood;"
c="a timorous capacity;" d="prudence;" e="circumspection;"
and f="infinite delay."

Having solved the syntactic puzzle, however, we discover
that we have passed by the whole point of the passage. Just
as the poet himself has "no skill to part" the elements from
one another, so he has intertwined them for us. By omitting
elements which would both semantically and syntactically
clarify the relationships of the parts to one another, he has
purposely brought together, graphically, nominals which
belong conceptually together as well. Again, the syntax

iconically mirrors the conceptual relationships.

71

3.3.6 DISCUSSION

Having examined a number of devices which Wordsworth
employs within his sentences, we are in a position to make
some general statements about the role of the sentence and
its constituents in Wordsworth's work. Primary among these,
of course, must be the two concepts which we have confronted
in most of the previous five sections: that Wordsworth's
sentence seeks to include a wide range of participants and
propositions, and that none of these seems to take greater
importance than the others. Permitting for a moment an
overly poetic metaphor, we might say that Wordsworth's
sentence is somewhat like the national cemetery at
Gettysburg: thousands of unique, individual persons lie
buried there together, and yet no one of them is of any
greater importance than the others. And the sum total of all
their deaths is but a single event in the course of the
"great civil war" (Lincoln, 1863).

Ward (1984) describes Wordsworth's writing thus:

Wordsworth is after all then expressing desire, that
nature be really unified in experience, so much so that
even literally disconnected things seem part of or close
to each other... Wordsworth's invitation is that we
share his longing to see, or make, a natural unity (pp.

70-71)

 

72

One way to formalize this use of the sentence as a
construction for unifying concepts is to suggest that any two
sentential constituents X and Y may be considered
linguistically closer (and thus, iconically, conceptually
closer) if they appear in a single sentential matrix,
separated by A B C constituents, as in (41), than if they

appear in separate sentences, as in (42):

(41) [ X A B C Y ]

(42)[X] [Y]

Numerous factors may mitigate the usefulness of such a
correspondence. For example, if X and Y are coreferential
noun phrases in (42), then they will certainly be extremely
close conceptually. On the other hand, if the same situation
exists in (41), then we may have apposition, and the
possibility that X and Y express the same participant within
the same proposition. This is conceivably "closer" than a
coreferential relationship in (42), but such matters would
probably always be highly context—dependent. In any case,
the preceeding discussion of Wordsworth's sentences has
served to establish some basis for positing (41) as
corresponding to a generally "closer" conceptual relationship
between X and Y than does (42), at least within Wordsworth's

work. A great deal of textual analysis (utilizing procedures

 

 

 

73

as yet undiscovered) would presumably be necessary to
validate this claim for linguistic systems in general. One
would also want to ask whether such generalizations could
apply to both written (premeditated) and oral (spontaneous)
discourse —- but these questions are beyond the scope of the
present work.

Having accepted at least partly the idea that X and Y in
(41) are "closer" than in (42), we may naturally want to
extend Haiman's (1985a) linguistic distance scale, presented
in 2.3, above, to include this new distinction (here, we will
use the typographical period symbol for a sentence break; #

indicates a word boundary; + indicates a morpheme boundary):

(43) Diminishing linguistic distance between X and Y

a. X . Y
b.X#A#B#Y
c. X # A # Y

d. X + A # Y

e. X # Y
f. X + Y
g. Z

In (43), the transition from a to b seems to allow a
rather large jump in the amount of material intervening
between X and Y. We might ask if boundaries intermediate to

those we have accounted for, such as those of phrases,

 

74

clauses, or even other marks of punctuation, such as commas,
dashes or semi—colons, should in fact be included between
items (43a) and (43b). And of course, we would predict that
they should. Again, fuller expansion of (43) is beyond the
scope of this investigation.

A similar question would be to ask if the same effect
holds at higher levels, as well. Does, for example,
inclusion in the same paragraph make two items likely to be
semantically closely linked?

The likely answer to this, it would seem, is "yes." The
paragraph is most certainly a thematic unit, and things
included in a single paragraph are, almost by definition,
somehow related to a single theme. Thus, it seems likely
(but certainly not guaranteed) that the same relatively
"close" relationship of X to Y will be implied by their
inclusion in a single paragraph.

Taken to an even higher level, this would suggest that
inclusion of X and Y within the same discourse (narrative,
poem, sitcom episode, etc.) implies that they are more likely
to be closely related than if they occur in totally different
discourses.

This last point is perhaps trivial, but it serves to
show an essential similarity between Haiman's idea of
"conceptual distance" and the long—recognized idea of
"theme." The two ideas are really not very different, in

that both recognize that concepts which are mentally

 

75

associated tend to be associated in linguistic expression, as
well. It is willingly admitted that this summary does
justice to neither idea, but it shows what they have in
common, which is really a great deal of what has made each of

them independently noteworthy.

4. SUMMARY AND IMPLICATIONS
. 4.0 SUMMARY

A functional View of natural languages presupposes the
ability for adaptive change over diachrony. Likewise,
adaptation presupposes some set of standards, internal or
external, which serve to discriminate between linguistic
structures, selecting, among possible forms, those which are
better-suited than others for performing various
communicative functions desired by language users. We have
adopted the stance that these "standards" are cognitive
strategies within the human mind.

Also presupposed by such a View is the idea that the
relationship between linguistic form and external function is
at least partially non—arbitrary. When observable form can
be related to performed or desired function in any consistent
way, we label the relationship between form and function
"iconic."

We have shown that iconicity involves linguistic
diagrams or images of mental concepts. We have further
suggested that iconicity, as a general principle of human
cognition, should be expected to be identifiable at all
levels of linguistic coding.

An attempt has been made to demonstrate the presence of
iconic principles in natural language syntax in a variety of
areas. Sequential ordering in the linguistic system often
corresponds to temporal sequential ordering. Likewise, the
occurrence of similar linguistic forms generally corresponds

to the performance of similar functions. Morphological

76

 

77

markedness may indicate semantic complexity. Children
demonstrate several speech tendencies which involve iconic
adaptation of adult language. Diachronic change may be
iconically motivated in several areas. Some syntactic
patterns demonstrate influence from cognitive
information-processing strategies. Finally, a correlation
may be identified between the size of a lexicon and the
presence of iconically motivated syntax. Together, these
arguments form a basis for believing that iconicity is widely
present in natural languages.

Given the likelihood that iconic principles may be
identified in human language, we have attempted to show their
presence in a specific corpus of William Wordsworth's poetry.
We have found that the phonological size of noun phrases
iconically indicates how topical their referents are within
the surrounding text. We have also seen that several iconic
principles may be employed to help in identifying the
grammatical relations of preverbally occurring objects.
Finally, we have shown that several characteristic features
of Wordsworth's sentences -- punctuation, subordination,
apposition, parallelism and ellipsis -- may serve as iconic
indicators of certain effects the poet is attempting to
achieve.

The goal of this thesis has been twofold: to
demonstrate the validity of iconic principles in linguistic

description, and to show the relevance of such principles to

78

an understanding of Wordsworth's poetry. It is hoped that
the reader will agree that progress has been made in both

areas .

4.1 GENERAL COMMENTS

For the most part, the discussion presented in this
thesis has been qualitative, although some quantitative study
has been presented in Section 3.1. The goal has been to show
that some of the rather impressionistic feelings a reader may
have about the effectiveness of Wordsworth's poetry may be
helpfully explained in terms of identifiable principles of
iconicity.

Labov has criticized the "overestimation" of functional
theorists who may argue strongly for a pervasive presence of
functional motivations in language without "paying attention
to all of the available data, rather than just those
utterances that favor the ideas under consideration"
(1987:314—15). In a study such as the present one, where
examples have obviously been carefully selected from the
corpus of data to demonstrate relevant points clearly, such a
criticism might be justified. This could only be the case,
however, if the purpose of the argumentation were to claim
that the iconic motivations proposed for explaining the
structures examined here could be found behind all (or even
many) other constructions, as well.

But no such claim has been advanced here. Rather,

 

79

there has been an attempt to show that iconic motivations are

clearly present in certain types of constructions, and that

 

in the cases of those constructions, wherever they may be
found across the entire corpus, iconic principles operate.
Generalization of the explanations offered here, either to
other constructions in Wordsworth's work or elsewhere in
human language generally, is certainly to be encouraged, but
caution is also advised; Labov's criticism may well be
justified if iconicity is bandied about as a catch-all
explanation for the observable forms of syntactic

constructions.

4.2 SYNTAX AND MEANING

Hill (1976) presents a fairly traditional interpretation
of the correlations between levels of language and their
function within various levels of literature. The
overlapping of "language proper" and "literature proper"
includes discourse phenomena —- "the area of sentences and
their relations to each other" (p.45) —- but not
sub-sentential areas such as morphology and syntax, which are

considered "non-literary." Table VI is reproduced from Hill:

 

80

TABLE VI
Levels of Language and Literature
from Hill (1976:47)

 

Level Language Literature Level
Uppermost Correspondence Cultural values Uppermost
nonlanguage meaning Structural and most nonliterary
analogical meaning
Correspondence
meaning
Language Structure and function Content structures Literature
proper of sentences, of literature proper
Discourse structure or Phonological
style structures of
literature (rhyme,
meter, etc.)
Structure and function Language as Lowest
of words, syntax material of nonliterary
Structure and function literature
of meaningful elements,
morphology
Structure and function
of sounds, phonology
Lowest Articulations and Irrelevant

nonlanguage

sounds

 

81

Chapter 3 has attempted to demonstrate identifiable
correspondences between synactic constructions and the
literary effects intended by Wordsworth. Such
correspondences have some bearing on the relationships of
elements of language to elements of literature.

Specifically, the identification of iconic constructions
in Wordsworth's syntax allows us to conclude that the "levels
of language" presented by Hill should not be considered
rigidly separated. The structure of noun phrases, the
identifiability of preverbal objects, and the expansive
sentential matrices which we have examined all help to
demonstrate how syntax may be used as a sort of "material of
literature" which corresponds to a higher level than just the
"lowest non-literary," as in Table VI. Rather, we have found
syntax capable of contributing meaning directly to the
discourse, by virtue of iconic relations it may bear to that
which is being discussed, or to the way in which the
discussion is presented or perceived. Thus, we should not
confine syntax to the lowest realms of "language proper," but
must recognize its potential direct interplay with semantic
information (Hill's "Correspondence meaning"), and with
specific literary effects which an author may seek,
consciously or unconsciously, to achieve (Hill's "Content
structures of literature").

We have seen only some initial ways in which such

correspondences may be identified; it is hoped that future

82

investigations of the sort undertaken here will be able to
increase our understanding of the role of syntax in

literature, as well as that of iconicity in syntax.

 

APPENDIX

APPENDIX
NOTES ON THE CORPUS

The corpus of Wordsworth's poetry utilized for this
study is somewhat eclectic, drawing on several periods and
genres from the poet's work. Provided here is a list of all
poems cited, with bibliographical references and brief notes
concerning each poem's significance.

The Excursion. Found in Wordsworth (1940:Vol.5:1-312).
The longest of Wordsworth's poems. In the preface to
its second edition, Wordsworth describes it as "a
philosophical poem, containing views of Man, Nature and
Society." It is introspective, with irregular meter.
It does not rhyme.

Home at Grasmere. Found in Wordsworth
(1896:Vol.VIII:235-57).
This work is in Wordsworth's meditative, introspective
style. It describes the poet's life at his boyhood
home.

The Idiot Boy. Found in Wordsworth (1940:Vol.2:67-80).
A narrative poem which utilizes both rhyme and fairly
rigid meter.

Michael: A Pastoral Poem. Found in Wordsworth
(1940:Vol.2:80-94).
Here, Wordsworth focuses specifically on a "pastoral"
scene, telling a story of the common man in his common
language.

Ode: Intimations of Immortality from Recollections of Early
Childhood. Found in Wordsworth (1896:Vol.VIII:189-98).
One of Wordsworth's most famous poems. Here, he
utilizes both rhyme and meter, in lines of unequal
length. The focus is meditative.

The Prelude or Growth of a Poet's Mind. Wordsworth (1926).
This epic in fourteen books is Wordsworth's second
longest, but greatest, work. It is entirely
self-revelatory, without regular meter or rhyme. For
this study, selections have been chosen from the
original (1805) version, although the poet later revised
it for publication in 1850.

83

B I BLI OGRAPHY

B I BL I OGRAPHY

Becker, Alton L. 1975. A linguistic image of nature: The
Burmese numeral classifier system. International
Journal for the Sociology of Language 5.109-21.

Brown, Cheryl. 1983. Topic Continuity in Written English
Narrative. In Giv6n, 1983, pp. 313-41.

Bybee, Joan. 1985. Diagrammatic iconicity in stem-inflected
relations. In Haiman, 1985b, pp. 11-47.

Comrie, Bernard. 1980. Language universals and linguistic
typology. Chicago: University of Chicago Press.

Dillon, George L. 1978. Language Processing and the Reading
of Literature. Bloomington: Indiana University Press.

Dixon, Robert M. W. 1971. A method of semantic description.
Semantics, ed. by Leon Jakobovits and Danny Steinberg,
436-71. Cambridge: University Press.

Givon, Talmy. 1983. (Ed.) TOpiC Continuity in Discourse: A
Quantitative Cross-Language Study. Amsterdam: John
Benjamins.

----- 1984. Syntax: a functional-typological introduction.

Vol. I. Amsterdam: John Benjamins.

----- 1985. Iconicity, isomorphism, and non-arbitrary coding

in syntax. In Haiman, 1985b, pp. 187-219.

----- 1987. Beyond Foreground and Background. In Russell 8.

Tomlin, ed. Coherence and Grounding in Discourse.
Typological Studies in Language, Vol. 11, pp. 175-88.
Amsterdam: John Benjamins.

Greenberg, Joseph H. 1985. Some iconic relationships among
place, time and discourse deixis. In Haiman, 1985b, pp.

271-87.

Greenbie, Marjorie. 1977. Wordsworth's Theory of Poetic

Diction. New York: AMS. Reprinted from Yale Studies in

English, #57, 1917.

Haiman, John. 1980. The iconicity of grammar: isomorphism and

motivation. Language 56.3:515~40.

----- 1983. Iconic and economic motivation. Language

59.4:781—819.
----- 1985a. Natural Syntax. Cambridge: University Press.

----- 1985b. (Ed.) Iconicity in syntax. Amsterdam: John

Benjamins.

84

85

Hetzron, Robert. 1976. On the Hungarian causative verb and
its syntax. The grammar of causative constructions, ed.
by Masayoshi Shibatani, 371-98. New York: Academic
Press.

Hill, Archibald A. 1976. Constituent and Pattern in Poetry.
Austin: The University of Texas Press.

Hopper, Paul J. and Sandra A. Thompson. 1985. The iconicity
of the universal categories "noun" and "verb." In
Haiman, 1985b, pp. 151—83.

Hyman, Larry and Karl Zimmer. 1976. Embedded tone in French.
Subject and topic, ed. by Charles Li, 191—211. New York:
Academic Press.

Jakobson, Roman. 1966. Implications of Language Universals
for Linguistics. Universals of Language, ed. by Joseph
Greenberg, 2nd ed., 263-78. Cambridge, MA: MIT Press.

Johnston, Ray. 1981. Conceptualizing in Nakanai and English.
Syntax and semantics in Papua New Guinea languages, ed.
by Karl Franklin, 212- 24. Ukarumpa, Papua new Guinea:
SIL.

Kirsner, Robert S. 1985. Iconicity and grammatical meaning.
In Haiman, 1985b, pp. 249-70.

Labov, W. 1973. The Boundaries of Words and their Meanings.
In C. N. Bailey and R. W. Shuy, eds. New Ways of
Analyzing Variation in English, pp. 340-73. Washington:
Georgetown University Press.

----- . 1987. The Overestimation of Functionalism. In Rene
Dirven and Vilem Fried, eds. Functionalism in
Linguistics, pp. 311- 32. Amsterdam: John Benjamins.

Lincoln, Abraham. 1863. Speech at the Dedication of the
National Cemetery at Gettysburg.

MacWhinney, B. 1973. How Hungarian Children Learn to Speak.
Unpublished Doctoral Dissertation, University of
California, Berkeley.

Payne, Thomas E. 1988 (Draft). Field manual for descriptive
linguistics.

Pierce, Charles S. 1932. Collected writings, 2: Elements of
logic. Cambridge, MA: Harvard University Press.

Prideaux, Gary. 1987. Processing Strategies: A
Psycholinguistic Neofunctionalism? In René Dirven and
Vilém Fried, eds. Functionalism in Linguistics, pp.
297—308. Amsterdam: John Benjamins.

86

Rehder, Robert. 1981. Wordsworth and the Beginnings of Modern
Poetry. Totowa, N.J.: Barnes & Noble.

Slobin, Dan I. 1985. The child as a linguistic icon—maker. In
Haiman, 1985b, pp. 221-48.

Smoczynska, M. 1985. Acquisition of Polish. In D.I. Slobin,
ed. The cross—linguistic study of language acquisition.
Hillsdale, NJ: Lawrence Erlbaum Associates.

Ward, J.P. 1984. Wordsworth's Language of Men. Totowa, N.J.:
Barnes & Noble.

Wordsworth, William. 1896. Poetical Works, Vols. I—XI. Ed.
by William Knight. London: Macmillan and Co.

----- . 1926. The Prelude. Ed. by E. de Selincourt. London:
Oxford University Press.

----- . 1940. Poetical Works, Vols. 1-5. Ed. by E. de
Selincourt. Oxford: Clarendon Press.