Wednesday, 20 June 2007

The question of form in relation to genre and 'genre categories'

I am now moving on to discuss this very thought provoking issue, namely form as a criterion of genre categorisation, and whether and how I should take this into account when selecting the data to analyse.

Surely, if by the criterion of function I ended up with some broad 'genre categories', if I apply the additional criterion of form I would be able to break down these categories further into genres (another criterion to use would be content), and this matters because genre ultimately is important because it determines our expectations and therefore how we receive a text or an interaction.

A good example from my data would be a comparison between the 'agony aunt' column and a lengthy text with an introduction (presenting and discussing the problem) and then a number of sections providing solutions to various aspects of the problem. The 'agony' column typically consists of an general 'blurb' (with phrases such as 'I am Samantha and I reply to all of your problems') and then a question apparently from a reader (presenting the problem) followed by 'Samantha's' answer (solution) and so on.

Arguably readers may take 'agony aunt' columns less seriously. They may read them for a laugh, or as shocking extreme situations (affairs, 'incompatible' couples, 'unusual' reasons for arguement in relationships etc.) and perhaps as somebody else's problem and nothing they themselves would need advice on. But what is the reason for that?

My take is that it is because of the position of the 'agony aunt' in the authority-friendliness scale, namely, that the advise provider does not have much authority in this case. 'Samantha' is a pseudonym, 'Samantha' is not a psychologist and has no other formal qualification, she does not quote statistics or research or other experts. Second, the only issues she deals with are very individual problems related to love life and does not make any attempt to claim that these may actually be widespread or include the reader - the tone is very personalised too.

Is any of the above a matter of form? Is, for example, the presence of statistics in other texts a formal characteristic, or a matter of content? How about the difference in the format (Q&A as opposed to Introduction + Sections)?

Well, my deep philosophical problem then is....

What is form?

However we define them, form, content/meaning and function are very deeply intertwined, as recognised by the cognitive and functional approaches to language. (I desperately need references on that).
I have no answer to the question 'what is form', but I would really like to know...
For my purposes (classifying rather than analysing texts) I chose to consider as 'form' anything that you can tell/recognise without necessarily understanding. The example that most illustrates this point is that of inflections indicating grammatical categories. In Greek there is a specific set of inflections a verb can have. If I know what these inflections are, I can tell whether any word is a verb even if I don't have the slightest idea what the word means. (For the same of the argument I ignore here that some morphemes identical in pronunciation and spelling can be noun endings etc.)
What would then be the 'form' of a whole text, as opposed to its meaning or function?
An obvious candidate is layout. This is not to say that the layout of a text does not have meaning and does not perform a specific function. It is, however, possible, to say with the first glance if a text is separated in sections, if it organised in a question and answer format (in the case where questions are represented as section titles, put in different fonts and separated from the answers with a blank line), if it has a bullet point format, if it is short and framed in a 'box' or table etc. It is also possible to tell certain texts formally marked on the top right or left hand corner as 'editorial', 'interview' etc. (although the latter does not necessarily help with genre categorisation, as many interviews are not marked as 'interviews', and some such marks/labels do not clearly indicate any specific genre, e.g. in my data there is the English label 'only 4 you' which includes a variety of genres).
But this alone doesn't say anything about genre categorisation, unless we further consider the specific function of these forms. Even a question is formally marked with a question mark, but its functions may range from requesting information to requesting some action on behalf of the hearer/reader or even to assert something (rhetorical questions) and, in my data, to indicate that something is a problem that requires discussion and solutions. We still have to acknolwedge the presence of formal generic characteristics, i.e. formats shared by texts of one genre, and then move on to see what these forms do.
Although I am saying that I am not considering form, I am however considering structure. This is a tricky one - is structure form, or is it content? Or is it primarily a functional characteristic of texts? I would say it is the way content is organised and put in order - but then isn't the question-answer format not the way content is organised?
In that by structure I mean 'underlying structure', and I specify this as an underlying discourse schema, this is not a formal characteristic in that it is not readily observable. In the problem-solution structure in the advice texts we might say that the fact that the problem is presented before the solutions is 'form'. But in order to find this one has to analyse the text and isolate the evidence indicating that something is represented as a problem. Moreover, the problem-solution structure has the function of justifying the existence of the whole text - you cannot have solutions without problems.
Insofar as I take structure into account, I could say that I am taking form into account. When it comes to layout, despite the fact that the organisation of the text on the page might be significant, I still categorise texts in 'genre categories' irrespective of layout. When it comes to the close analysis of the 'advice' category, the genre membership of each text may come up as significant, but then again I suspect that layout will rather be symptomatic of function/ communicative purpose. So I would not venture to draw any conclusions or make any big claims about how the texts are preceived/ received and what their 'hidden' communicative purposes are before a close analysis.

Wednesday, 13 June 2007

My paper on genre and data selection

I posted the handout of my paper earlier on. It deals with the methodological issue of data selection, which is never as interesting as discussing findings, but nevertheless I got some very good comments (both good/positive and good/constructive - thanks to the people who commented).

The main issue is this: I have collected 6 different magazines, 3 copies of each. 18 volumes in total. I have to select which texts to analyse from each magazine, and the sample from each magazine has to be comparable across the magazines. What better way to do it than select specific genres to analyse, genres present in all magazines such as the editorial, the 'agony aunt', or the advertisements?

Well I thought that was too restrictive. We all can tell which text the editorial is (it even says so on the top of the page) - but there are texts in the magazines which don't fall into a 'folk categorisation' because people simply don't care enough to stick them in a category. Does this mean that nobody will ever analyse those? Or should I as an analyst decide what genres they belong to? Or should I use a category other than 'genre' for my classification?

Well I ended up with the four broad 'genre categories' or 'function categories' I mention in the handout:
- Promotional
- Advice
- Commentary
- Gossip

(This idea is the outcome of a conversation with Costas Gabrielatos, thanks Costa!)

These categories may consist of different genres. E.g. the 'advice' category includes the 'agony aunt', interviews with doctors, nutritionists or cosmetologists giving advice on health, nutrition or skincare, small bulleted lists, long texts etc.

Comments I got related to my presentation:
1. I neglect form
2. I shouldn't be neglecting form, because it determines the way we see and treat texts (and consequently plays a role in the effect of texts, which is what we are interested in ultimately)
3. Should I be saying that only 'promotion' and 'advice' texts perform a 'directive' speech act? Are (not) all texts in lifestyle magazines 'directive'?

One of my points has been that I am not interested in form but in function, and this is my criterion of categorisation. The question then arises: should I be neglecting form so much? (This was both a comment from the conference and from my supervision meetings before that). To this I have a practical answer and some impractical further questions:

First of all, I am not denying that form is not important for determining genre membership. But my purpose was not a taxonomy of the genres within magazines per se, but to find a way to construct a body of texts which make sense to analyse. In some way. And if I was going to select as a dataset of texts which should share function AND form, this would be too restrictive.

BUT this does not mean that because I ended up with a convenient categorisation (although this is open to further questioning, as in Comment 3 above) I should (appear to) say that in principle genres or forms don't matter, or any such folly. And I probably I SHOULD actually end up classifying the texts within the 'function categories' in genres, although now I am at liberty to include in my anylisis text of unidentified/unidentifiable genre identification as long as I have identified their broad function.

So I'm saying that at this stage what is more important is the fact that the texts in the categories share similarities in terms of function (advice) and the textual devices through which this is expressed, than the variation in form.

But still. Is it? This leads to point 2: If forms determine the way we see texts and the effect of texts on us, have I been missing the wood for the tree? If I want my data selection to be well-motivated, and if I am interested in ideology, shouldn't I be concerned with putting together texts with the same form, belonging to the same genre and consequently having the same kind of effect? (Maybe I am misinterpreting the point of this comment, need to check but for the time being I will pretend I know what I am talking about).

I have to finish this post now but I will continue it soon.


I am finally posting on my blog. It is supposed to be an 'academic' content blog, but rather than being isolated in my 'glass tower' I have been rather to absorbed by down-to-earth stuff to sit down and write anything in here.

I have had some inspiration and want to write some things, but I now realise they will make no sense to anyone but me unless I give some background. Maybe I will at some point :) The directly relevant background of my inspiration is the Birmigham PhD Critical Discourse Analysis thing called The First Aston University Postgraduate Conference in English Language and Linguistics. I am copy pasting my handout below and will post my thoughts and comments in a separate post. It's long and boring and looks bad, but hey maybe someone will find the reference list useful :) (or some of my friends may want to quote me but don't remember the reference, hehe)

Genre-based data selection and classification for Critical Discourse Analysis

The First Aston University Postgraduate Conference in English Language and Linguistics
Birmigham, U.K., 05/06/2007

Contextualising this paper in my research

Data: Greek lifestyle magazines
- Women’s (Cosmopolitan, Madame Figaro, Marie Claire)
- Men’s (Nitro, Playboy, Status)

Research question: Exploring the manifestation of gender ideologies in the data

Narrowing down focus:
- Selecting titles to analyse
- Selecting certain genres to analyse which will be present across magazines à comparable samples

Magazines are heterogeneous, comprising of a variety of different kinds of texts, which would require different methods of analysis if I was going to analyse them from cover to cover.

Various kinds of texts in magazines are not always clearly identified/identifiable genres.

What is ‘genre’?

- A kind/ type of text
- A discourse type (Fairclough, 1992)

Swales’ definition of genre:
a class of communicative events, the members of which share some set of communicative purposes… In addition to purpose, exemplars of a genre exhibit various patterns of similarity in terms of structure, style, content and intended audience (1990: 58).

Critical Linguistics and Critical Discourse Analysis link genres as communicative events to the (conventional(ised)) social activities achieved through (conventional) means – genres (Bakhtin, 1986: 60; Kress, 1985/1989: 19; Fairclough, 1992; 2001; Wodak, 2001: 66).

Functional approaches to language emphasise genres as
- Functions
- Realisations of communicative purposes
- Means of achieving social actions
- ‘Speech acts’

In short, genres are
‘how things get done, when language is used to accomplish them’ (Martin, 1985: 250)

And the criterion for assigning genre membership to a text is ‘what does the text do’?


‘what does it appear to be doing?’

Cf. Askehave (1999): overt vs. ‘hidden’ communicative purposes

Categorisation of texts in my data according to function:

- Promotion
- Advice

- Commentary
- Gossip

Ø These categories are too broad to be considered genres – each ‘genre’ ends up too heterogeneous. E.g. ‘promotional’ as discussed in Bhatia (1993) ends up including brochures, advertisements, book blurbs and job application cover letters (Askehave, 1999).
These are ‘genre categories’ (or should I say 'function categories') - despite significant differences in form and/or content, their overall communicative purpose is roughly the same. The analysis shows that they also share a number of linguistic means towards this communicative purpose. They consist of texts varying in length, content and format, which one could break down in different genres.

Ø Openness of categories restricted by the fact that they occur within the same super-genre ‘lifestyle magazine’ and could be seen as sharing ‘situational context’ as well as ‘co-text’ (which is why the categories are not too heterogeneous).

Ø Genre identification depends on our purposes (Chandler, 1997) -
for CDA we may want to focus on this level, since broad functions of texts cut across formal characteristics, and are linked to ideology (cf. Martin, 1992).

Ø Genres, super-genres (e.g. lifestyle magazines, academic journals etc.), and genre categories have a radial structure (Rosch, 1973; Lakoff, 1987), including more prototypical and more marginal members. They share generic characteristics through family resemblances, i.e. not all texts will have all of the generic characteristics (Swales, 1990).

How to identify texts belonging to a certain genre/ genre category?

Ø Paradox:

To take a genre such as the 'western', analyse it, and list its principal characteristics, is to beg the question that we must first isolate the body of films which are 'westerns'. But they can only be isolated on the basis of the 'principal characteristics' which can only be discovered from the films themselves after they have been isolated (Tudor, 1974: 135, cited in Gledhill, 1999: 138).

-> In order to select data for analysis, one has to analyse the texts before selecting them!

Ø Extra-textual cues to aid with genre classification (Swales, 1990; Bhatia, 1993; Askehave and Swales, 2001):
Background knowledge of discourse community using the genre, social context in which the discourse community is using the genre, literature and insiders/ experts from within the discourse community

Ø Too intuitive. Genre identification is often glossed over in order to move on with the actual analysis, whether it is called ‘genre analysis’ or not (e.g. Swales, 1990; 2004, Bhatia, 1993; Doctor-patient interaction in West, 1984; Fairclough, 1992: Ch. 6; Wodak, 1996: Ch. 2). This is because academics analysing the genres in question are members of the relevant discourse communities already.

Ø In the case of lifestyle magazines, members of the discourse community are not very concerned with the various genres present in lifestyle magazines, and do not have a ‘folk categorisation’ for all of them.

Ø Preliminary analysis: first on a more macro-level (overall schematic structure), then on a more micro-level.

Focus on ‘advice’ genre category in Greek women’s lifestyle magazines

Ø Underlying structure (Discourse schema, van Dijk, 1985):
Problem – Solution/ Question - Answer
Problem (elaboration) – suggested action (elaboration) – solution/ outcome of action (elaboration)
Lexis such as: problem, issue, tip, solution, advice, be careful etc.

Ø Overall speech act performed: Directive

Ø Style: Mixture of authority and solidarity (à ‘advice’ rather than ‘command’ or ‘request’)

- Distant
- Categorical modality, ‘absolute truths’
- High deontic modality (must, should)
- Scientific or quasi-scientific jargon
- Reference to ‘people who have the problem’ in 3rd person plural
- Authority quoted – exclusive we

- Friendly, familiar
- Deontic modality of various degrees (more hedging through expressions such as ‘you may want to…’ – negative politeness BUT less authority)
- Informal vocabulary
- Addressing the reader (2nd person singular, questions, imperatives)
- Inclusive we

Ø Advice texts more reader- oriented than other genre categories.
Commentary & gossip: About something/ someone else
Promotional: directive, too, thus addressing the reader more, but also about the commercial items promoted
Advice: about YOU

Ø Topics vary:
Health (nutrition, exercise, alcohol consumption, STDs), Appearance (Hair, Skincare), Career, Money management, Relationships (with men, friends, colleagues), Sex etc.

Ø Forms vary too:
Small tables/ bullet points
Long solid text
Long text including sections, one for each problem/question (Section title can also be in form of question)
Interviews with experts
Readers’ letters answered by authority (Question & Answer format)

Examples from data

See Texts A, B and C at the end of the handout (I will post them later).

Concluding remarks

Ø In critical research genres are of interest in terms of function (rather than form).
Ø Depending on the social issue in question, the data and the functions we are interested in, functions can cut across genres.
Ø We can rely on textual elements for a preliminary categorisation of texts according to function.
Ø We need further deeper analysis to examine further social functions, ‘hidden’ purposes and ideologies, but we need a sound method of initially selecting our corpus.


ASKEHAVE, I. (1999) Communicative Purpose as Genre Determinant. Journal of Linguistics, 23.
ASKEHAVE, I. & SWALES, J. M. (2001) Genre identification and communicative purpose: a problem and a possible solution. Applied Linguistics, 22, 195-212.
BAKHTIN, M. M. (1986) The problem of speech genres. Speech Genres and Other Late Essays [transl. Vern W. McGee; edited by Caryl Emerson and Michael Holquist]. Austin, University of Texas Press.
BHATIA, V. K. (1993) Analysing Genre: Language Use in Professional Settings, London; New York, Longman.
CHANDLER, D. (1997) An introduction to genre theory.
FAIRCLOUGH, N. (1992) Discourse and Social Change, Cambridge, Polity Press.
FROW, J. (2005) Genre, London; New York, Routledge.
GLEDHILL, C. (1999) Genre. IN COOK, P. & BERNINK, M. (Eds.) The Cinema Book. 2nd ed. London, British Film Institute.
KRESS, G. (1985/1989) Linguistic Processes in Sociocultural Practice, Oxford, Oxford University Press.
MARTIN, J. R. (1985) Process and text: two aspects of human semiosis. IN BENSON, J. D. & GREAVES, W. S. (Eds.) Systemic Perspectives on Discourse, Vol. 1. Norwood, NJ, Ablex.
LAKOFF, G. (1987) Women, Fire and Dangerous Things: What Categories Reveal About the Mind, Chicago, Chicago University Press.
SWALES, J. M. (1990) Genre Analysis: English in Academic and Research Settings, Cambridge, Cambridge University Press.
SWALES, J. M. (2004) Research Genres: Exploration and Applications, Cambridge, Cambridge University Press.
TUDOR, A. (1974) Theories of Film, London, Secker and Warburg/ BFI.
VAN DIJK, T. (2001a) Discourse, ideology and context. Folia Linguistica, XXV, 11-40.
VAN DIJK, T. (2001b) Multidisciplinary CDA. IN WODAK, R. & MEYER, M. (Eds.) Methods of Critical Discourse Analysis. London; Thousand Oaks; New Delhi, Sage.
VAN DIJK, T. (2006) Discourse, context and cognition. Discourse Studies, 8, 159-177.
VAN DIJK, T. A. (1985) Structures of news in the press. IN VAN DIJK, T. A. (Ed.) Discourse and Communication: New Approaches to the Analysis of Mass Media Discourse and Communication. New York, W. de Gruyter.
WEST, C. (1984) Routine Complications: Troubles with Talk between Doctors and Patients, Bloomington Indiana University Press
WODAK, R. (1996) Disorders of Discourse, London, Longman.
WODAK, R. (2001) What CDA is about - a summary of its history, important concepts and its developments. IN WODAK, R. & MEYER, M. (Eds.) Methods of Critical Discourse Analysis. London, Sage.