Department of Languages
University of Jyväskylä, Finland
28 June 2007
Linguistic interviews can be considered a well-established
genre – or perhaps a coordinated set of several different genres. The purpose
of the interviews is to extract linguistic data from the informants, and
data
here has meant primarily low-level linguistic information on
phonetics, phonology, morphology, and somewhat also on syntax, of different
languages and language variants. The semantic, pragmatic and discoursal
points of view were for a long time totally neglected.
The grand rule of the interviewers was to say as little as possible and let the informant speak. Long, uninterrupted sequences of speech were the thing. The interviewer was only to guide the informant to a topic where his or her speech would flow most naturally.
To put it in another way, the purpose of the interviews didn’t pose any
demands on the topic, or subject, of the discussion. In principle, the
interviewer was free to follow his or her own feelings on this. In practice,
however, the individual interviewers followed well-established conventions
in this. When linguistic interviews were introduced to the discipline, one of
the main motives for them was to record old speech
, older dialects
that were about to die out. That’s why the informants were as a rule selected
from older people, and preferably from those who had lived all their lives in
one and the same dialect area. Also, topics of the interviews were chosen so
that the old speech
would be manifested as much as possible: tales of
old times, reminiscences of old forms of living, and life story narratives
were much called for.
However, no topic could guarantee success. Interviewees are always
different from one another, and while some could talk almost endlessly on
olden times
, some just couldn’t. When the possibilities of the topic
were exhausted, or if staying on the topic didn’t ensure those long
uninterrupted speech sequences by the informant, it was up to the interviewer
to find out a new line of inquiry.
The first task for the interviewer is to find out how to signal (rather than tell) the informant that the only thing we really are looking for are those long uninterrupted sequences of speech. Experienced interviewer could accomplish a lot by clever extralinguistic signals such as nodding their heads a lot: instead of saying or even mumbling anything, they just nodded encouragingly. But before the informant can continue anything, he or she has to get started.
Life story narratives were one way of getting just there. Either by asking
direct yes/no questions (Did you like school?
) or more roundabout open
questions (What did you used to be doing after school in those days?
),
or even attempting to assert something on the informant (like I guess you
were X then
), the informant just could get enough material to
work on on his own.
But be it as it may, the big picture of linguistic interviews is always a mix of different topics, different ways of dealing with those topics, and quite probably, of different (sub)genres.
Now, I’ll the term narrative to refer to those discourse sequences where (a) the informant goes well beyond the question, or lead, of the interviewer: says more than absolutely necessary in order to be cooperative; and (b) where time, or changes in time, are involved. Further, life story narratives are specifically about the informant himself, his changing along time. (I admit these are not exact definitions; I’m merely setting these up as working terms.)
Even a casual browsing through the transliterations of the interviews makes it immediately clear that there’s no specific narrative mode, or genre, that would differ once and for all from all the other modes or genres used during the interview. That is, given two sequences of discourse, one cannot say which is narrative and which isn’t looking at the textual features alone.
Let’s look at this closer.
The first linguistic level to consider is the lexical level: the words or rather semantic word fields used in discourse.
The lexicon is of course intimately tied with the topic of the discourse. Thus, it varies widely from speaker to speaker, as the life stories of each speaker are different from each other.
Since the purpose of the interview, from the point of view of the
interviewer, is not to extract information
from the informant, that
is, the topic is quite irrelevant to the interviewer, it is changed as soon
as it begins to look like it’s not going anywhere – the informant has
nothing more to say and is expecting the interviewer to carry on.
Further, the dialect interviews typically change the topic of the
discussion in a seemingly haphazard way. Even if the grand total of life
stories
remains, life is a big subject. For example, one informant talked
quite a lot about his favourite authors, his working as an accountant in a
grocery store and his relative being a diplomat – widely differing topics all
in all! Another one was supposedly more reluctant to ignite
, so he was
asked questions about houses and apartments in general, the types of houses
and apartment he’d lived in and would like to live in more specifically, and
the city traffic.
So, even just these two informants would give a very variable lexicon for
the life story narrative
genre, if we were looking for one.
Next level down, we have syntax; but syntactic research has hardly ever been carried with real-life corpuses. So we come to the next level still, the morphology.
When looking at the morphological features of speech, the difference between
narrative and non-narrative is even more opaque than in the lexical features.
Since the outside
– physical and social – situation is the determining
factor in morphological variation, hardly anything can be said to change when
switching between the narrative and non-narrative phases of discourse.
There are certain rare cases where grammar reflects genre. In Finnish, one
might be the specific form of genitive plural case, the so-called second form
of genitive plural being used often in formal or archaic (old-fashioned)
discourse. Narrative could be a borderline case, being more formal and
somewhat more archaic
(in a loose sense) than the non-narrative parts
of interviews. However, life story narratives are again probably the most
informal mode of narrating! so the grammar switching
phenomenon would
probably get cancelled anyway. Be that as it may, I haven’t found even one
example of the narrative
use of 2nd genitive plural in the dialect
interviews in my corpus.
Lowest down on the linguistic scale, is phonology. Here
we have what might be loosely described as the dialectal
features
of speech.
The corpus I have used is from a sociolinguistic research that was carried in here in Tampere: it contains speech from the same informants from years 1977 and 1997, 20 years apart (some new speakers were added in 1997, though). Its main use has been to study how (and if) individual features speech change over time.
However, from the point of view of my present objective, all the speakers are from Tampere and manifest – although in highly variable ways – the local speech patterns. There’s variance but not enough to warrant a conclusion that phonological features were being used in genre distinctions.
But here we do have an interesting possibility. Since narrative means,
effectively, stepping out of the timeframe and setting of the current situation,
it could be signalized using phonological cues. Let’s say the narrative contains
several different characters, different people. These people could be
represented in the speech so that their individual voices
were
represented by some dialectal features of their speech.
In my present material, I have yet to find cases of this; but I have seen (or rather, heard) it being used in more traditional dialect interviews. There it really could be the case that phonology were used as a cue for narrative mode.
The point is, it would be a cue. No more, no less.
Below phonology, and according to most linguistic theories, already outside linguistics, there are phonetic features of speech.
Now I’m referring to features of speech that carry no linguistic (that is, distinctive) meaning such as voice quality, some aspects of pitch, different registers of voice etc.; most of these can have linguistic significance in some languages, but none actually has in Finnish.
It’s a well-known phenomenon in conversational analysis that people can use different kinds of voice when for example, mocking or ironizing something or trying to imitate someone else. As in phonology, this could get a narrative function.
To conclude, there are no special lexical, morphological, phonological, or phonetic features that would immediately and without context reveal a given sequence of speech as narrative or non-narrative. An unpromising prospect? I wouldn’t say so.
According to my own theory of genres (to be presented in my PhD thesis:
Nieminen, forthcoming), this is only to be expected. Genres do not set up
distinctive features
– neither formally nor semantically. Genre is not
so much recognized
from texts as it is overlaid
on them. That
is, genres are not found
inside texts: it is put there intentionally
for purposes of discourse.
If something is (or isn’t) narrative, it is not because of its textual
overall features but because we were either told by the speaker
or surmised ourselves that it’s narrative. Something being narrative
has a functional purport for the discourse: it gives something to our
interpretation of it.
So, looking at the totality of the text isn’t the answer.
Distributive features seem to lead us nowhere. Instead – and this is my
hypothesis now, we should look at the specific points in time when speaker
metatextualizes
something in order to make us understand that what
follows is narrative (or whatever).
In other words, even if it’s not systematically signalized when being inside a particular genre, the points of genre change are quite often signalized. Not necessarily always, though. Sometimes what happens happens outside the text we can see from the transliteration – in the situational context that has been already lost. Also, there are conventions governing discourse. Even if something isn’t present physically, it may be very much present socially: even more so.
Furthermore, we should never forget that interviews form a dialogue. A coordinated and perhaps one-sided dialogue, but a dialogue nonetheless. Since the interviewer is heavily trying to coax the informant into giving us his life story, or any other narrative for that matter!, we should also be closely looking at the interviewer’s discourse. What happens inside an interview is a joint project.
So we need a theoretical apparatus with which to model this punctual-wise nature of genre action. The most promising linguistic candidate I have found is the contextualization theory.
Contextualization theory is a relatively recent theory in linguistics: it originated in anthropological linguistics and the sociology of speech (or, sociology of communication, both terms have been used interchangeably) in the late 1970’s and early 1980’s. In the beginning, as formulated by John Gumperz in his 1982 monograph Discourse Strategies, the concept of conceptualization was little more than a way of going around the strict delimitation between text and context: contextualization referred only to the textual cues that the speaker used to point out the correct interpretational context of the text.
Lately, however, anthropological linguists such as Richard Bauman, Charles Briggs and William Hanks (cf. Hanks 1989, 1996; Bauman & Briggs 1992; Bauman 1992), have used the term contextualization in a way that’s much more interesting both theoretically and in practice. Briefly, to them, contextualization means a two-way process of getting things out of their original context and denoting the intended interpretative context in new settings.
One of the key concepts is decontextualization, the ability of texts to rise out of their immediate contexts of forming, or as it were, to form autonomous or semi-independent objects that seem to be able to be interpreted without the original context. Language is the main decontextualizer, but other convention, such as genres, may do the trick as well.
The other key concept of the theory is recontextualization
which means the necessity of always contextualizing, that is, putting into a
new context, anything that was decontextualized, when it’s being actually
used. There can be no communication without a context. But
recontextualization does not just depend on the outside
context: the
inside
context of the text, the information that was previously
decontextualized into it, acquires now a new function. The text carries
information about its intended context inside itself.
Also, recontextualization means that the practice of creating discourse can be considered as a multi-level operation: on one level, there is the current discourse, on another, there is contextualization: metatextual operations that the speaker uses to point out the relevant context of the discourse.
The main point is: contextualizing a text into a genre does not need distributive genre features in the text. It will be enough that at key points of discourse the hearer is reminded of the relevant context – in this case, the relevant genre.
My empirical findings are yet very preliminary. I’m rather still trying to connect the theoretical (or even, metatheoretical) model of my PhD thesis to a very different kind of textual corpus, the dialect interviews. However, I think that even these preliminary findings show that distributive genre features are, and have always been, a blind alley. Rather, we need to look closer at the key points in discourse where genres are signalled; and in here, the theory of contextualization will get very useful.