THE SUSANNE CORPUS
Release 1, 6th September 1992
Geoffrey Sampson
School of Cognitive & Computing Sciences
University of Sussex
Falmer, Brighton BN1 9QH, England
geoffs@uk.ac.susx.cogs
INTRODUCTION
The SUSANNE Corpus has been created, with the sponsorship of the Economic
and Social Research Council (UK), as part of the process of developing
a comprehensive NLP-oriented taxonomy and annotation scheme for the
(logical
and surface) grammar of English.[1] The SUSANNE scheme attempts to
provide
a method of representing all aspects of English grammar which are
sufficiently
definite to be susceptible of formal annotation, with the categories and
boundaries between categories specified in sufficient detail that,
ideally,
two analysts independently annotating the same text and referring to the
same scheme must produce the same structural analysis. The SUSANNE
scheme
may be likened to a "Linnaean taxonomy" of the grammatical domain: its
aim (comparable to that of Linnaeus's eighteenth-century taxonomy for the
domain of botany) is not to identify categories which are theoretically
optimal or which necessarily reflect the psychological organization of
speakers'
linguistic competence, but simply to offer a scheme of categories and
ways
of applying them that make it practical for NLP researchers to register
everything that occurs in real-life usage systematically and
unambiguously,
and for researchers at different sites to exchange empirical grammatical
data
without misunderstandings over local uses of analytic terminology. On
reasons why such a scheme is needed at the present juncture in NLP
research,
see e.g. Sampson (1992, forthcoming).
Note that a sharp distinction is drawn here between the terms "scheme"
and
"system". A "parsing scheme", or "analytic scheme", refers to a range of
notations and guidelines for using them which prescribe to a human
analyst
what the appropriate grammatical annotation for a language example should
be. A parsing "system" on the other hand refers to a software system
which
automatically produces analyses (according to some parsing scheme) of
input
language examples. A parsing scheme defines the target which a parsing
system
hits (or misses). The SUSANNE Corpus represents part of the definition
of
a parsing scheme. It has been produced largely manually, not as the
output
of an automatic parsing system.
The SUSANNE analytic scheme is defined in detail in a book by myself,
ENGLISH FOR THE COMPUTER, forthcoming from Oxford University Press.
The Chairman of the Analysis and Interpretation Working Group of the
US/EC-sponsored Text Encoding Initiative has proposed its adoption as
a recognised TEI standard. The SUSANNE scheme aims to specify annotation
norms
for the modern English language; it does not cover other languages,
although it
is hoped that the general principles of the SUSANNE scheme may prove
helpful in developing comparable taxonomies for these.
Regrettably, Release 1 of the SUSANNE Corpus is not a "TEI-conformant"
resource, though aspects of the annotation scheme have been decided in
such a way as to facilitate a move to TEI conformance in later releases.
The working timetable of the Initiative meant that relevant aspects
of the TEI Guidelines were not yet complete at the point when
the SUSANNE Corpus was ready for initial release; delaying this release
would have been unfortunate.
The brief description of the SUSANNE Corpus which follows cannot replace
the
very detailed statements to be found in ENGLISH FOR THE COMPUTER, and any
user aiming to do serious work with the Corpus or its annotation scheme
would need to consult the book. Nevertheless, it may be useful to have
a summary statement included with the electronic Corpus.
The present SUSANNE annotation scheme originated in work carried out by
myself in collaboration with Professor Geoffrey Leech FBA and others in
the
years 1983-85 to produce a database of manually analysed sentences from
the
LOB Corpus of written British English, as a source of statistics for
probabilistic automatic-parsing techniques; this database, which has not
been (and will not now be) published, is described in Garside et al.
(1987: ch. 7). The annotation scheme of this "Lancaster-Leeds Treebank"
represented surface grammar only, without indications of logical form.
It
subsequently seemed desirable to extend this scheme to include methods
for
representing logical grammar, and to refine both surface and logical
aspects
of the annotation scheme by applying it to a larger body of texts. The
only
way that a parsing scheme can in practice be made increasingly adequate
is
in the way that the English Common Law develops, by collecting and
systematizing the body of precedents generated through detailed
consideration
of more and more individual cases that arise in real life. Accordingly,
Project SUSANNE took a subset of the Brown Corpus of written American
English which had been manually analysed by Alvar Elleg<aring>rd's
group at Gothenburg (Elleg<aring>rd 1978), and reworked the annotations
in this under-used resource in order to turn them into a scheme
consistent
with that used in the Lancaster-Leeds Treebank but including
specifications
of logical as well as surface structure: several categories of
information
not indicated in either Lancaster-Leeds or Gothenburg schemes were
also added.[2] (On Brown and LOB Corpora, see e.g.
Garside et al. (1987: 4-5).)
The finished SUSANNE parsing scheme has thus been developed on the basis
of
samples of both British and American English. It is oriented chiefly
towards
written language; however, on another project sponsored by the Royal
Signals
and Radar Establishment[3] my team produced extensions to the SUSANNE
scheme
for annotating the distinctive grammatical phenomena of spoken English,
and these extensions are specified in ENGLISH FOR THE COMPUTER (though
they
are not used in the SUSANNE Corpus and are not discussed further here).
It should be noted also that the scheme has emerged through a process of
detailed critical discussion of analytic standards by some ten people
over a decade; apart from myself, the leading role in the early years of
these discussons was taken by Geoffrey Leech, whose standing as an
English
grammarian needs no emphasis.
The SUSANNE Corpus itself comprises an approximately 128,000-word subset
of the Brown Corpus of American English, annotated in accordance with the
SUSANNE scheme. The original motives for producing this database
included
that of providing better statistics for probabilistic parsing; but in
this
respect Project SUSANNE was overtaken after its inception by projects
(notably
Mitchell Marcus's Pennsylvania Treebank project, cf. Marcus & Santorini
(forthcoming)) which have used quasi-industrial methods to generate far
larger bodies of grammatically-analysed material. However, the SUSANNE
scheme
may be unparallelled in the extent to which its categories have been
refined
and tested through detailed consideration of the almost endless small
quirks
of the texts to which they have been applied, and in the degree of
precision to
which the resulting guidelines for using the categories have been
documented
-- thus defining analytic standards which permit annotation of future
material
to be extremely self-consistent. Accordingly the SUSANNE Corpus is
offered
to the research community primarily as a demonstration of the application
of the parsing scheme, evidencing the fact that the scheme has survived
the
test of experience rather than being a merely aprioristic system. The
SUSANNE Corpus functions, as it were, like a collection of type specimens
appended to a botanical taxonomy.
Although the accompanying first release of the SUSANNE Corpus has
undergone
considerable proof-checking, it unquestionably still contains many
errors.
I intend to correct these in future releases; I shall be extremely
grateful
if users discovering errors will log these and send details to me,
preferably
by post rather than e-mail.
STRUCTURE OF THE CORPUS
The SUSANNE Corpus consists of 64 files (apart from this documentation
file),
each containing an annotated version of one 2000+ word text from the
Brown
Corpus. Files average about 83 kilobytes in size,
thus the entire Corpus totals about 5.3 megabytes.
The file names are those of the respective Brown texts, e.g. A01, N18.
Sixteen texts are drawn from each of the following Brown genre
categories:
A
G
J
N
press reportage
belles lettres, biography, memoirs
learned (mainly scientific and technical) writing
adventure and Western fiction
The Corpus thus samples each of the four broad genre groups established
on the basis of word-frequency data by Hofland & Johansson (1982: 27).[4]
Each file has a line (terminating in a newline character) for each "word"
of
the original text; but "words" for SUSANNE purposes are often smaller
than
words in the ordinary orthographic sense, for instance punctuation marks
and the apostrophe-s suffix are treated as separate words and assigned
lines
of their own. (For details on the rules by which orthographic words
are segmented, as well as on all other analytic matters mentioned below,
see ENGLISH FOR THE COMPUTER.)
Each line of a SUSANNE file has six fields separated by tabs (that is,
there is one tab after each of fields 1 to 5, but a newline after field
6).
Each field on every line contains at least one character.
The six fields on each line are:
1
2
3
4
5
6
reference
status
wordtag
word
lemma
parse
Apart from the tab and newline characters used to structure fields and
records, all bytes in each of the 64 SUSANNE files are drawn from a
subset of
the 94 graphic character allocations of the International Reference
Version
("IRV") of ISO 646:1983 "Information Processing -- ISO 7-bit coded
character
set for information interchange", from hexadecimal 21 (exclamation mark)
to hex 7E (tilde). These codes are assumed for SUSANNE purposes to
represent the graphic symbols assigned by the IRV system.
Twelve members of the IRV character set are not used in the Corpus,
namely
(all codes hexadecimal):
23
24
27
2F
5C
5E
5F
60
7B
7C
7D
7E
gate
generalized currency unit
prime
solidus
reverse solidus
circumflex
underline
grave
opening curly bracket
vertical bar
closing curly bracket
tilde
The space character, hex 20, which is classified by ISO 646 as a control
code
also does not occur in the SUSANNE Corpus.
Where text characters cannot be adequately represented directly within
the
resulting 82-member character set, they are represented by entity names
within angle brackets. Where possible these are drawn from Appendix D
to ISO 8879:1986, "Information Processing -- Text & Office Systems -Standard Generalized Markup Language (SGML)". For instance, "<eacute>"
stands for lower-case "e" with acute accent. Symbols in angle brackets
are
used also to represent such things as typographical shifts, which for
purposes of grammatical analysis are conveniently represented as items
within the word-sequence: e.g. "<bital>" stands for "begin italics".
REFERENCE FIELD
The reference field contains nine bytes which give each line a reference
number that is unique across the SUSANNE Corpus, e.g. "N06:1530t".
The first three bytes (here N06) are the file name; the fourth byte is
always a colon; bytes 5 to 8 (here 1530) are the number of the line in
the "Bergen I" version of the Brown Corpus on which the relevant
word appears (Brown line numbers normally increment in tens, with
occasional
odd numbers interpolated); and the ninth byte is a lower-case letter
differentiating successive words that appear on the same Brown line.
(SUSANNE lines are lettered continuously from "a", omitting "l" and "o".)
STATUS FIELD
The status field contains one byte. The letters "A" and "S" show that
the
word is an "abbreviation" or "symbol", respectively, as defined by Brown
Corpus codes (Francis & Ku<ccaron>era 1989: 12).
The letter "E" shows that the word is (or is part of) a misprint or
solecism in the original text (details are logged in ENGLISH FOR
THE COMPUTER). On the great majority of lines, to which none of these
three categories apply, the status field contains a hyphen character.
WORDTAG FIELD
The SUSANNE wordtag set is based on the "Lancaster" tagset listed in
Garside
et al. (1987: Appendix B); additional grammatical distinctions have been
drawn in this set, and these are indicated by suffixing lower-case
letters
to the Lancaster tags. For instance, "revealing" is tagged "VVG"
(present
participle of verb) in the Lancaster scheme, but as "VVGt" (present
participle of transitive verb) in the SUSANNE scheme. Apart from the
lower-case extensions, the wordtags are normally identical to the
Lancaster
tags: punctuation marks are assigned alphabetical tags beginning Y...
(e.g. YC for comma), and the dollar sign which appears in some Lancaster
tags for genitive words is replaced by G (e.g. GG for the apostrophe-s
suffix), so that the modified Lancaster tags always consist wholly of
alphanumeric characters, beginning with two capital letters. (In a few
cases, tags from the Lancaster set have been merged or eliminated from
the SUSANNE scheme in the light of experience.)
The tag YG appears in the wordtag field to represent a "trace" -- the
logical
position of a constituent which has been shifted elsewhere, or deleted,
in the surface grammatical structure.
The SUSANNE tagset comprises 352 distinct wordtags, not counting tags for
elements of "grammatical idioms" (see below); a few of these wordtags
never occur in the SUSANNE Corpus. The wordtags are listed, and their
application rigorously defined, in ENGLISH FOR THE COMPUTER -- in the
case
of closed wordclasses, by enumeration of their members, and in the case
of open classes by rules for choice between alternative tags. These
rules
refer to information in a specified published dictionary (the OXFORD
ADVANCED LEARNER'S DICTONARY OF CURRENT ENGLISH, 3rd edition).
WORD FIELD
The word field contains a segment of the text, often coinciding with a
word
in the orthographic sense but sometimes, as noted above, including only
part of an orthographic word. In general the word field represents all
and
only those typographical distinctions in the original documents which
were recorded in the Brown Corpus (Francis & Ku<ccaron>era 1989: 10-15),
though
in certain cases the SUSANNE Corpus has gone behind the Brown Corpus to
reconstruct typographical details omitted from Brown.
Certain characters have special meanings in the wordfield, as follows:
+
(occurs only as first byte of the wordfield) shows that the
contents of the field were not separated in the original text
from the immediately-preceding text segment by whitespace (e.g. in
the
case of a punctuation mark, or part of a hyphenated sequence split
over
successive SUSANNE lines);
-
the line corresponds to no text material (it represents the "trace"
for a grammatically-moved element);
<...> enclose entity names for special typographical features, as
discussed
above, either taken from ISO 8879:1986 Appendix D or created for
the
SUSANNE Corpus -- for instance "<pand>" stands for "either plus
sign or
ampersand", since the Brown Corpus makes no distinction between
these
characters.
LEMMA FIELD
The lemma field shows the dictionary headword of which the text word is
a form: the field shows base forms for words which are inflected in
the text, and eliminates typographical variations (such as sentenceinitial capitalization) which are not inherent to the word but relate to
its use in context. (In the case of "words" to which the dictionary-form
concept is inappropriate, e.g. numerals and punctuation marks, the lemma
field contains a hyphen.) Orthographic forms in the lemma field are
those
of a specified dictionary (the OXFORD ADVANCED LEARNER'S DICTIONARY OF
CURRENT ENGLISH, 3rd edition).
Project SUSANNE aimed also to indicate the senses which polysemous words
bear
in context, via codes relating word-tokens to numbered subsenses in a
specified dictionary. The book ENGLISH FOR THE COMPUTER provides a
detailed
coding scheme for representing this information. Unfortunately, this
aspect
of the project's output proved to contain a number of inadequacies, and
the information does not appear in Release 1 of the Corpus. It is hoped
to include it in later releases.
PARSE FIELD
The contents of the sixth field represent the central raison d'<ecirc>tre
of the SUSANNE Corpus. They code the grammatical structure of texts as
a sequence of labelled trees, having a leaf node for each Corpus line.
Each text is treated as a sequence of "paragraphs" separated by
"headings".
A "paragraph" normally coincides with an ordinary orthographic paragraph;
a "heading" may consist of actual verbal material, or may be merely a
typographical paragraph division, symbolized "<minbrk>" in the word
field.
Conceptually, the structure of each paragraph or heading is a labelled
tree with root node labelled "O" ("Oh" for a heading), and with a leaf
node
labelled with a wordtag for each SUSANNE word or trace, i.e. each line of
the Corpus. There will commonly be many intermediate labelled nodes.
Such a tree is represented as a bracketed string in the ordinary way,
with
the labels of nonterminal nodes written "inside" both opening and closing
brackets (that is, to the right of opening brackets and to the left of
closing brackets). This bracketed string is then adapted as follows for
inclusion in successive SUSANNE parse fields. Wherever an opening
bracket
immediate follows a closing bracket, the string is segmented, yielding
one
segment per leaf node; and within each such segment, the sequence
opening-bracket + wordtag + closing-bracket, representing the leaf node,
is replaced by full stop. Thus each parse field contains exactly one
full
stop, corresponding to a terminal node labelled with the contents of
the wordtag field, sometimes preceded by labelled opening bracket(s) and
sometimes followed by labelled closing bracket(s), corresponding to
higher
tagmas which begin or end with the word on the line in question.
Brackets are square except in the case of nodes immediately dominating
the
"trace" wordtag YG, which are represented with angle brackets.
Nonterminal node labels in the SUSANNE scheme contain up to three types
of
information: a FORMTAG, a FUNCTIONTAG, and an INDEX, in that order.
In a label containing a formtag and one or both of the other two
elements,
a colon separates the formtag from the other elements. A functiontag is
always a single alphabetic character, and an index is a sequence of three
digits; restrictions on valid combinations of elements within a node
label
mean that complex labels can always be unambiguously decomposed into
their
elements.
RANKS OF CONSTITUENT
Apart from nodes immediately dominating traces, all node have labels
including formtags, which identify the internal properties of the word or
word-sequence dominated by the node. The shape of a parse-tree is
defined
in terms of a hierarchy of formtag ranks:
1
wordlevel formtags (begin with two capital letters; formtags
of all other ranks begin with one capital and contain no further
capitals)
2
phraselevel formtags (begin with one of:
N V J R P D M G)
3
clauselevel formtags (begin with one of:
S F T Z L A W)
4
rootlevel formtags (begin with one of:
O Q I)
Each grammatical clause, whether consisting of one or more words, is
given
a node labelled with a clauselevel formtag. Each immediate constituent
of
a clause, whether there are one or more such constituents and whether the
constituent consists of one or more words, is given a node labelled with
a phraselevel formtag, unless the constituent belongs to a wordlevel
category that has no corresponding phraselevel category
(e.g. punctuation marks, conjunctions), or to a rootlevel
category (e.g. a direct quotation, formtagged Q). Thus a clause
consisting
of one verb will be assigned a clauselevel formtag (e.g. Tg for presentparticiple clause) which singularily dominates a phraselevel formtag
(e.g. Vg for "verb group beginning with present participle") which in
turn
singularily dominates a wordlevel formtag (e.g. VVGi for "present
participle
of intransitive verb").
Other than by these rules, and in certain other limited circumstances
specified in ENGLISH FOR THE COMPUTER, singulary branching does not
occur.
An intermediate phraselevel node is inserted between a higher phraselevel
node and a sequence of words dominated by it only if two or more of those
words form a coherent constituent within the higher phrase. A clause
which
fills a slot standardly filled by a phrase (e.g. a nominal clause as
subject
or object) will not have a phrase node above the clause node unless the
clause proper is preceded and/or followed by modifying elements that are
not part of the clause.
Detailed rules for deciding constituency in various debatable cases,
for placing items such as punctuation marks within parse trees, etc.
are laid down in ENGLISH FOR THE COMPUTER.
FUNCTIONTAGS AND INDICES
Functiontags, identifying roles such as surface subject, logical object,
time adjunct, are assigned to all immediate constituents of clauses,
except for their verb-group heads and certain other constituents for
which
function labelling is inappropriate.
Indices are assigned to pairs of nodes to show referential identity
between
items which are in certain defined grammatical relationships to one
another.
For instance, a phrase raised out of a lower clause to act as object in
a higher clause, as in "John expected Mary to admit it",
will be assigned an index identical to that assigned to the trace showing
the logical position of the item in the lower clause. The (artificial)
example quoted would be represented as:
[Nns:s John] expected [Nns:O999 Mary] [Ti:o <s999 TRACE> to admit [Ni:o
it]]
-- where the index 999 shows that the trace acting as logical subject
(symbolized s) of the "admit" clause is coreferential with "Mary" which
acts as surface object (O) of the "expected" clause; the logical object
(o)
of the "expected" clause being the infinitival subordinate clause (Ti).
In some cases, movement rules displace a constituent into a tagma within
which it has no grammatical role (for instance, an adverb which is
logically a clause constituent may interrupt the verb group -- sequence
of auxiliary verbs and main verb -- of the clause): in such cases
the functiontag is G ("guest"). Constituents which do not logically
belong
below the node which immediately dominates them in surface structure
are always given G functiontags and indices linking them to their logical
position. With that exception (and with one other exception not
discussed
here relating to co-ordination), functiontagging is used only for
immediate constituents of clauses.
ENGLISH FOR THE COMPUTER lists the categories of surface/logical-grammar
discordance which are represented by the SUSANNE scheme, and the approved
methods of representing them. The SUSANNE analysis is always chosen so
as to be as far as possible neutral as between alternative linguistic
theories.
THE FORMTAGS
The SUSANNE formtags are as follows:
Rootlevel Formtags
O
Oh
Ot
Q
I
Iq
Iu
paragraph
heading
title (e.g. of book)
quotation
interpolation
tag question
scientific citation
Clauselevel Formtags
S
Ss
Fa
Fn
Fr
Ff
Fc
Tg
Ti
Tn
Tf
Tb
Tq
Z
L
A
W
main clause
quoting clause embedded within quotation
adverbial clause
nominal clause
relative clause
"fused" relative
comparative clause
present participle clause
infinitival clause
past participle clause
"for-to" clause
"bare" nonfinite clause
infinitival relative clause
reduced ("whiz-deleted") relative clause
other verbless clause
special "as" clause
"with" clause
Phraselevel Formtags
N
V
J
R
P
D
M
G
noun phrase
verb group
adjective phrase
adverb phrase
prepositional phrase
determiner phrase
numeral phrase
genitive phrase
The various phrase categories take lower-case subcategory symbols
which can be combined in any meaningful combination (e.g. the verb group
"must have been noticed" would be formtagged "Vcfp"). The phrase
subcategories are:
Vo
Vr
Vm
Va
Vs
Vz
Vw
operator section of verb group, when separated from remainder
of V e.g. by subject-auxiliary inversion
remainder of V from which Vo has been separated
V beginning with "am"
V beginning with "are"
V beginning with "was"
V beginning with other 3rd-singular verb
V beginning with "were"
Vj
Vd
Vi
Vg
Vn
Vc
Vk
Ve
Vf
Vu
Vp
Vb
Vx
Vt
V beginning with "be"
V beginning with past tense
infinitival V
V beginning with present participle
V beginning with past participle
V beginning with modal
V containing emphatic DO
negative V
perfective V
progressive V
passive V
V ending with BE
V lacking main verb
catenative V
Nq
Nv
Ne
Ny
Ni
Nj
Nn
Nu
Na
No
Ns
Np
"wh-" N
"wh...ever" N
"I/me" head
"you" head
"it" head
adjective head
proper name
unit noun head
marked as subject
marked as nonsubject
singular N
plural N
Jq
Jv
Jx
Jr
Jh
"wh-" J
"wh...ever" J
measured absolute J
measured comparative J
postmodified J
Rq
Rv
Rx
Rr
Rs
Rw
"wh-" R
"wh...ever" R
measured absolute R
measured comparative R
adverb conducive to asyndeton
quasi-nominal adverb
Po
Pb
Pq
Pv
"of" phrase
"by" phrase
"wh-" P
"wh...ever" P
Dq
Dv
Ds
Dp
"wh-" D
"wh...ever" D
singular D
plural D
Ms
M headed by "one"
NON-ALPHANUMERIC FORMTAG SUFFIXES
Formtags may also contain non-alphanumeric symbols, including:
?
*
%
!
"
interrogative clause
imperative clause
subjunctive clause
exclamatory clause or other item
vocative item
Other non-alphanumeric symbols represent co-ordination structure. Under
the SUSANNE scheme, second and subsequent conjuncts in a co-ordination
are analysed as subordinate to the first conjunct; thus a co-ordination
of the form:
chi, psi, and omega
(whatever the grammatical rank of the word-sequences chi, psi, etc.)
would
be assigned a structure of the form:
[chi, [psi], [and omega]]
The formtag of the entire co-ordination is determined by the
properties of the first conjunct (except for singular/plural
subcategories
in the case of phrase categories to which these apply); the later
conjuncts (which will often be transformationally reduced) have nodes
of their own whose formtags mark them as "subordinate conjuncts". The
following symbols relate to co-ordination (and apposition) structure:
+
@
&
subordinate conjunct introduced by conjunction
subordinate conjunct not introduced by conjunction
appositional element
co-ordinate structure acting as first conjunct within a
higher co-ordination (marked in certain cases only)
Co-ordination is recognised as occurring between words as well as between
higher-rank tagmas. Therefore nonterminal nodes may have formtags
consisting
of wordtags followed by co-ordination symbols, thus (using "WT" to stand
for an arbitrary wordtag):
WT&
WT+
WT-
co-ordination of words
conjunct within wordlevel co-ordination that is introduced
by a conjunction
conjunct within wordlevel co-ordination not introduced by a
conjunction
(A wordlevel co-ordination always takes an ampersand on its formtag;
phrase or clause co-ordinations do so only in very restricted
circumstances.)
Also, certain sequences of orthographic words, in certain uses, are
regarded as functioning grammatically as single words ("grammatical
idioms").
For instance, "none the less" would normally be treated as a grammatical
idiom, equivalent to an adverb (for which the wordtag is RR). In such
cases, the nonterminal node dominating the sequence has a formtag
consisting of an equals sign suffixed to the corresponding wordtag;
and the individual words composing the grammatical idiom are not
wordtagged
in their own right, but receive tags with numerical suffixes reflecting
their membership of an idiom. (The sequence "none the less" would be
formtagged RR=, and the words "none", "the", and "less" in this context
would be wordtagged RR31 RR32 RR33.) ENGLISH FOR THE COMPUTER includes
exhaustive listings of closed-class grammatical idioms.
Note that formtags of the forms WT& WT+ WT- WT= rank as wordlevel
formtags
for the purposes of determining tree structure as discussed above.
THE FUNCTIONTAGS
Functiontags divide into COMPLEMENT and ADJUNCT tags: broadly, a given
complement tag can occur at most once in any clause, but a clause may
contain
multiple adjuncts of the same type. The scheme of adjunct categories has
been developed from the classification of Quirk et al. (1985).
Complement Functiontags
s
o
S
O
i
u
e
j
a
n
z
x
G
logical subject
logical direct object
surface (and not logical) subject
surface (and not logical) direct object
indirect object
prepositional object
predicate complement of subject
predicate complement of object
agent of passive
particle of phrasal verb
complement of catenative
relative clause having higher clause as antecedent
"guest" having no grammatical role within its tagma
Adjunct Functiontags
p
q
t
h
m
place
direction
time
manner or degree
modality
c
r
w
k
b
contingency
respect
comitative
benefactive
absolute
Detailed guidelines for the application of these functional categories is
included in ENGLISH FOR THE COMPUTER.
NOTES
[1]
The support of the Economic and Social Research Council (ESRC)
is gratefully acknowledged. Project SUSANNE, "Construction of an
Analysed Corpus of English", was funded by ESRC award
no. R00023 1142, over the period 1988 to 1992. "SUSANNE" stands for
"Surface and underlying structural analyses of naturalistic English".
I should like to express my warmest thanks to the team who worked on
Project
SUSANNE, namely Robin Haigh, H<eacute>l<egrave>ne Knight, Tim Willis,
and Nancy Glaister.
[2]
I thank Alvar Elleg<aring>rd for permission to circulate a research
resource derived from the work of his group.
[3]
Ministry of Defence research contract no.
D/ER/1/9/4/2062/151(RSRE),
"A Speech-Oriented Stochastic Parser".
[4]
The Brown texts included in the Gothenburg and hence the SUSANNE
Corpora are as follows:
A01
A09
G01
G09
J01
J09
N01
N09
A02
A10
G02
G10
J02
J10
N02
N10
A03
A11
G03
G11
J03
J12
N03
N11
A04
A12
G04
G12
J04
J17
N04
N12
A05
A13
G05
G13
J05
J21
N05
N13
A06
A14
G06
G17
J06
J22
N06
N14
A07
A19
G07
G18
J07
J23
N07
N15
A08
A20
G08
G22
J08
J24
N08
N18
For details on the source publications from which these texts were
sampled,
see Francis & Ku<ccaron>era (1989).
REFERENCES
A. Elleg<aring>rd (1978) The Syntactic Structure of English Texts.
Gothenburg Studies in English, 43.
W.N. Francis & H. Ku<ccaron>era (1989) Manual of Information to
Accompany a
Standard Corpus of Present-Day Edited American English, for use
with Digital Computers (corrected and revised edition).
Department of Linguistics, Brown University, Providence, Rhode
Island.
R.G. Garside, G.N. Leech, & G.R. Sampson, eds.
Analysis of English. Longman.
K. Hofland & S. Johansson
American
English. Longman.
(1982)
(1987)
The Computational
Word Frequencies in British and
M.P. Marcus & Beatrice Santorini (forthcoming) "Building very large
natural language corpora: the Penn Treebank". To appear in
N. Ostler, ed., Proceedings of the 1992 Pisa Symposium on
European Textual Corpora.
R. Quirk, S. Greenbaum, G. Leech, & J. Svartvik (1985)
Grammar of the English Language. Longman.
A Comprehensive
G.R. Sampson (1992) "Probabilistic parsing". In J. Svartvik, ed.,
Directions in Corpus Linguistics: Proceedings of Nobel Symposium
82,
Mouton de Gruyter.
G.R. Sampson (forthcoming) "The need for grammatical stocktaking". To
appear in N. Ostler, ed., Proceedings of the 1992 Pisa Symposium on
European Textual Corpora.
SUSANNE.doc