Introduction + Cohesion

Thu Jan 21 15:14:41 UTC 1999

Hello,

My name is Sanna-Kaisa Tanskanen; I am a research fellow of the Academy
of Finland at the Department of English, University of Turku, Finland.
I'm currently working on my dissertation, which deals with the effect of
communicative conditions on the use of lexical cohesion in spoken and
written English (in face-to-face conversation, prepared speech,
mailing-list language and academic writing, more specifically). My other
interests include discourse features of computer-mediated communication,
interactional and collaborative features of monologue, and Early-Modern
English letter-writing considered from a discourse perspective.

I joined the list the day before yesterday, and as I was skimming through
the archive files to get an idea of what had already been discussed on
the list (and how), I noticed there were a few messages on cohesion. I
hope you will excuse me for continuing a little on this topic.

I think the discussion was initiated by James Cornish's question to the
Hallidayans on the list. Let me stress that I am not a Hallidayan,
although a lot of what I do has been greatly influenced by him and also
by some other systemic functionalists. I feel that at the present stage
of discourse-oriented language studies, still very much characterised by
emergent ideas, models and theories as well as highly fluctuating
terminologies, the best way to find interesting answers may not be to
follow any one model but to try and combine several approaches, which
will usually reveal that they have more similarities than differences
regardless of for instance their possibly very different terminology.
This is not to say that following one particular model would not yield
interesting results as well.

But back to James's question of whether or not in empirical studies of
texts the types of cohesion in Halliday & Hasan (1976) can be compressed
into types of lexical cohesion, as suggested by Hoey (1991). Although
James's question is primarily practical, theoretical ramifications are
unavoidable. First I think it should be clarified that Hoey does not
actually propose that all types of cohesion should be compressed into
types of lexical cohesion, let alone "reduced", as suggested by Zouhair
Maalej. Hoey notes that the majority of the types of cohesion in his
material (and in Halliday and Hasan's example texts and probably in any
longer text) are instances of lexical cohesion (as in Halliday & Hasan,
lexical is here opposed to grammatical), and thus the space allotted to
the treatment of lexical cohesion in H & H does not seem to correspond at
all to its actual importance in texts. It should, however, be noted that
H & H (1976, p.287) are themselves aware that lexical cohesion (and
especially their category of collocation) needs to be further analysed and
more strictly defined. The objective of Hoey's book is, to put it briefly,
to try to highlight the importance of lexical cohesion, which he decides
to do by concentrating on repetition (repetition understood very broadly
as 'saying something again').

However, in classifying the relations which he considers instances of
lexical cohesion (or repetition), Hoey notes that there are also other
devices than purely lexical ones serving the same function, i.e.
repeating something: the pronoun systems and substitutes (one, do and
so). Although these are grammatical members of closed systems, their
function is to substitute for lexical items, and this is why Hoey
includes them in his analysis. The issue of whether or not to treat these
items in the same way as lexical items is basically very simple: which is
considered more important, their difference in form or their similarity
in function?

It is probably evident by now that pronouns and substitution items are
included in my own analyses as well. -- I must add, however, that my
model of analysis differs from Hoey's in two respects. First, even more
than Hoey does in his analysis, I avoid using general lexical semantic
terms for lexical relations, and follow McCarthy (1988) by adopting a
discourse-specific approach. In other words, I do not start from
ready-made lexical semantic classifications, but from a text, and then
try to establish which items are related in that particular text.
However, although a discourse-specific and a lexical semantic approach
are distinct, they are not irreconcilable; I use non-lexical-semantic
terms, for instance the term equivalence instead of synonymy, to draw
attention to the fact that the justification for a lexical relation should
be sought from the text in which the related items occur, not from a
decontextualised classification. In most cases a semanticist would no
doubt agree with me on the validity of the relations, but not necessarily
in all of them. Secondly, my analysis includes collocation relations
(collocation as defined by Halliday & Hasan); I have further defined the
category on the basis of the work by e.g. Jordan (1984) and Martin (1992).
-- Recognising pronoun repetition and pronouns substituting for lexical
items naturally means that the number of repetition relations per text is
higher than it would be without them, and the cohesive picture therefore
also more accurate, I think. As to the importance of the actual
substitution items (50 pages in Halliday & Hasan; lexical cohesion is
covered in 20 pages...), I could note here that in my material they do
not seem to be very frequent; it could perhaps be said that they are
locally  important in the texts, but not globally. They are more frequent
(or less infrequent) in face-to-face conversations than in my other
texts, and very rare indeed in academic writing. Prepared speech and
mailing-list language are very similar as regards the number of
substitution items and would be situated in between conversations and
academic writing on this "dimension". Following e.g. Biber, I do not
think there can be a simple dichotomy between spoken and written language;
what I try to show in my dissertation is that the use of cohesion varies
according to the demands of the communicative conditions in which the
texts have been produced and processed.

I'm afraid this message is getting rather lengthy, so I'd better sign off
now before I truly get carried away. Needless to say, I would be more
than willing to discuss cohesion and coherence in more detail (publicly
or privately), and I'm looking forward to interesting discussions on
other topics as well.

I list below two papers by Hoey which I think have not been mentioned; the
first one is a 'patterns of lexis' analysis on narrative texts, which also
introduces one further repetition category, namely closed set; the second
one discusses intertextual aspects of lexical cohesion.

Hoey, M. 1994. "Patterns of lexis in narrative: a preliminary study", in
Tanskanen, S-K. & B. Warvik (eds). 1994. Topics and Comments: Papers from
the Discourse Project. (Anglicana Turkuensia, 13). Turku: University of
Turku. 1-40.

Hoey, M. 1995. "The lexical nature of intertextuality: a preliminary
study", in Warvik, B., S-K. Tanskanen & R. Hiltunen (eds). 1995.
Organization in Discourse: Proceedings from the Turku Conference.
(Anglicana Turkuensia, 14). Turku: University of Turku. 73-94.

Anglicana Turkuensia web page:
http://www.utu.fi/hum/engfil/publicat.html

* * *
Sanna-Kaisa Tanskanen
Department of English
University of Turku
20014 Turku, Finland
sakata at utu.fi