I first began writing this post in the early Fall, but set it aside due to time constraints and extreme frustration. I wanted some time for better perspective on the issues. SBL has come and gone, I have had some great opportunity to interact with some practitioners. In rereading the post and cleaning up a few things, I think it is finally time to post this. Be aware that my comments are primarily focused on this applying Systemic Functional Linguistics within biblical studies, and primarily to those applying it to syntax. I would appreciate hearing comments from those of you with an opinion.
Original post:
This is a follow up to a previous post on the need for reading the sources of those whom you are reading. In other words, instead of citing X on Y, read Y. There comes a practical limit in research to which this can be done. But folks, there is nothing more critical to successful and useful research than a sound and robust theoretical framework. I cannot emphasize this enough. This is also a follow-up to a series of posts culminating here on why I am not a systemic functional linguist. In the name of practicing what I preach, and in prep for a paper on the verbal aspect of the historical present at SBL this year, I hit the library.
To date, most of what I know about Systemic Functional Linguistics has come from those that are applying it to biblical languages, especially Greek. I have spent the last few days reading full-fledged linguists who are plying their trade with SFL. It has been interesting to say the least.
I am coming to the conclusion that those applying the Hallidayan model in NT studies may not be true Hallidayans, but more like an offshoot. I read an interview with Halliday here and found it quite helpful in understanding what he is after. He is very theoretically driven to solve some problems that will probably not help me much in my exegesis, but they are legitimate questions. What’s more, I have yet to find an SFLer outside of biblical studies using a symmetrical approach to markedness. It would seem that Halliday would use a symmetrical cline to represent some of his probabilistic data mentioned in the interview, but in skimming through his Introduction to Functional Grammar (2004) they are not to be found. Within NT studies, markedness clines form the foundation of most work, be it word-order, aspect, morphology, you name it. Outside, they are proving difficult to find.
For those of you interested in applying linguistics and discourse analysis to biblical studies, I would offer two pieces of advice.
- Read the original sources that are cited by those applying them to NT studies. It is relatively easy to misunderstand someone when you start out, and such mistakes can compound if not corrected. Go read Halliday on Halliday to form an opinion, read folks that have made linguistics their primary profession. I encourage you to do the same with my material. If I cite Lambrect on information structure or mental representations, then by all means read Lambrect. It will definitely be worth your time.
- Read competing theories or approaches that offer alternative explanations to the problem you are trying to solve. In my case of information structure, Lambrecht was my chosen model. However, Functional Grammar, Relevance Theory, Role and Reference Grammar, and SFL all offer competing explanations. In doing my reading, I found that I could reconcile the basic frameworks into a unified account, with the exception of Halliday (see chapter 9 of Discourse Grammar of the GNT).
One thing that I did find that was consistent with Hallidayans both within and without biblical studies: the propensity to resignify common terminology to serve their purposes. For instance, Halliday begins with the Prague School concepts of Theme and Rheme. The theme is ‘what the sentence is about’, and rheme, is ‘what is said about it’. Simple enough, right? This is fundamentally what the approaches above use as their core concept, though cognitive and other constraints have been brought to bear to add precision. The core idea remains unchanged, though. Not so with Halliday.
Halliday (2004:64) defines theme as “the element which serves as the point of departure of the message; it is that with which the clause is concerned.” He then goes on to associate Theme with initial position in English. This causes a problem, in that there are constructions that allow one to place Rheme into the initial position in a clause. One reason for making this change is to make the info stand out, to “emphasize” it. The most common construction for this is the it-cleft, e.g. “It is by grace that you have been saved” as in Eph 2:8. The translators (virtually all) use an it-cleft to reflect the marked word order in Greek, which emphasizes the word grace. However, Halliday’s system classifies this position as Thematic, even though it contradicts the Prague School notions of Theme and Rheme. So what is the solution to this problem? How is it resolved? There seem to be two preferred solutions to such dilemmas in SFL:
- Add another layer of analysis, in this case the “Information unit.” This allows you to maintain your original model without having to change anything. Don’t change, just add. Halliday makes it work for him.
- If you must change, then change the meaning of the original concept that caused the dilemma, not your own notion. In this case, Theme is changed from what the Slavic “Founding Fathers” intended to what Halliday needed to make the SFL model work. “Just play through,” I always say.
I wish I was making this stuff up, I really do, but there are just too many such examples in SFL. Other methods might do this as well, but not with the consistency or pervasiveness I have seen in SFL.
So why are many of the SFL applications to Koine Greek so complex? Part of the reason in my opinion is Solution 1 above. They have added completely unnecessary complexity to already complicated issues, all in the name of preserving the original model. Language is complex enough as it is. Solution 2 adds another unnecessary complexity to the process, since it is much harder to reconcile what is being claimed by Kwong’s use of SFL with Lambrecht, Dik, Van Valin or some other competing theory.
One final observation. SFLers seem bent on perpetuating the model against all competitors, yet there is little to no interaction with the competitors. In reading Halliday on information structure, there is no engagement of competing theories, no effort to plot his proposal in the larger constellation of the theoretical universe. In contrast, the first third of Lambrecht is devoted to interaction with the field. Hallidayans are intent on using a purely SFL framework, which seemingly results in reading primarily Halliday or his followers. I have seen the same tendency in Relevance Theory by relying upon Sperber and Wilson, but there is still a fair degree of interaction with competing theories.
If you want to invest the time, money and effort to learn Halliday, if you think that his concepts hold the answers to the complexities of language and are chuck full of useful nuggets that will help with exegesis, then go for it. I think there are much simpler and more productive ways of accomplishing most of these tasks. SFL provides a productive framework for preliminary sketches in new areas of study, but areas like information structure in Koine and Classical Greek are well beyond this point. Studies like Nick Bailey’s really drive that point home.
BTW, if you use SFL and can respond to the issues I raise, I would love to hear from you.
Hi Steve,
I have been following your posts with interest, and was aware that
you were saying things about SFL which were not true of core SFL, but
seem to be true of SFL in biblical studies (e.g., the markenesss issue).
However, last time this site would not let me post, I hope this time
different.
Some points:
a) You are right saying most in SFL do not address those in other
theories. It is a problem. However, in some cases, the work of Halliday
predates that of the others (except Prague School, which he does acknowledge.
For instance, on Theme, Halliday first published on this in 1967.
Functional Grammar, Relevance Theory, Role and Reference Grammar
all arose later. Equally so with Halliday and Hasan’s work on
Cohesion in English, where they were first, and later works were
derivative.
b) Introduction to Functional Grammar does not cover markedness
clines. True. The function of the book is a “simple” introduction
to grammar for university students of English. Markedness clines
would probably confuse them. (Hasan likes to say that there
is not such think as marked options – each option is the appropriate
choice for particular contexts).
Actually, Halliday would not use the term (I think) ‘markedness clines’.
What he does argue is that some choices in systems are equal valued
(e.g., choice between past and present tense, while others are skewed
(e.g., active passive).
However, there are clearly problems with this: it is a central
notion in SFL that the probabilities of linguistic choices are
very much in relation to the context of situation of the text,
which Halliday seems to forget in his writing of sytemic probabilities.
In any case, I am happy that you have come to see that SOME of the
problems in biblical SFL re not due to SFL itself, but some practicioner’s
own contribution.
Mick O’Donnell
Mick,
Thanks for taking the time to post the comment. I have been reflecting on the post myself for the last day, wondering whether to pull it. Here a couple points. First, I do not think that Halliday is attempting to answer the kinds of questions that I am in the area of syntax and pragmatics. He touches on these areas, but his interest is English and the configurational choices available to the speaker. I think this may explain the stark difference between his approach and that of the other major theories I mentioned.
Second, I have read quite a bit of the SFL literature, particularly those applying Halliday’s model. There seems to be great interest in Halliday in South America. These practitioners are the ones that I find engaging the opposing theories in the literature, but then not resolving the core differences other than to change the terminology.
Regarding Halliday predating other information structure models, his ‘Transivity’ articles primarily sketched the outline of his proposed model, but did not develop syntax much until the 80s from what I can tell. The work of Firbas clearly predates this, and forms the basis for most of the other functional approaches. Givon and Dik had full models by the early 80s, though both have continued to modify them in response to problems that arise from new data and applications. The rise of RT has had a significant impact as well, forcing linguists to incorporate some kind of cognitive element into their model. I would expect the same from Halliday, regardless of whether he was first or not.
Fundamentally Halliday’s work on cohesion and coherence are helpful and worthwhile. I think there are easier ways of reaching the same conclusions, but to each his own. My greatest frustration is the application of SFL to areas for which they are ill-suited. I think for those pursuing a Brown and Yule style analysis, some of Halliday’s work would be a great resource. I had not realized how far the practitioners in NT had moved from Halliday himself. Ultimately, it is these practitioners that I am addressing in my research, and their incarnation of SFL. If they are misapplying the model, I do not have the time or interest to learn Halliday to correct their work. As I stated earlier, SFL offers little for the kinds of questions I am trying to answer, particularly in the cross-linguistic context of Bible translation. I will do some reworking of the post to differentiate SFL proper from NT.
Regarding “markedness clines” I am simply trying to describe the implementation, and do not expect Halliday to use the term. Halliday himself seems far more interested in the asymmetrical system of the decisions than in a prominence-based ranking of semantic weight. This is good to know.
I find it interesting that Fawcett of Cardiff is working to revise and simplify the SFL framework, at least according to some recent articles by Butler on “focus” and information structure. I began looking into Fawcett’s work, but it still seems very much an English-based model of syntax. If you can suggest sound, representative treatments applying SFL to a non-configurational language like Greek, I will gladly read it. Most of what I have found regards Spanish, and is what generated my comment about changing the terminology where the framework does not fit, rather than addressing possible flaws in the framework itself.
Again, thanks for the helpful feedback. I will be shifting to directly addressing NT practitioners from here on out.
Again, thanks for offering the clarification.
I’m not going to comment on the post itself because I haven’t had time to process it fully yet and Mick’s reply is very thorough and useful. I would like to say that I think these sorts of discussions are excellent and I was very pleased to stumble across your blog. The only person I am aware of working on Greek and SFL is Astika Kappagoda.
Lisa, I am coming to realize more clearly that the brand of SFL being applied in Koine Greek likely has little resemblance to what you are doing in Cardiff. It would really take reading an article or two for you to see what I mean. My primary background is in pragmatics and information structure, working primarily along the lines of Knud Lambrecht for the latter. I would suggest the following:
Porter Stanley E. – O’Donnell Matthew Brook, «The Greek Verbal Network Viewed from a Probabilistic Standpoint: An Exercise in Hallidayan Linguistics».
Thanks for the comment, and I hope for a better distinction to be made between what I have come to know about SFL proper and it permutation in New Testament studies.