Transcribing Conversation
Introduction
In order to analyze conversation (or indeed any talk-in-interaction) the audio-recorded data is typically transcribed into a written form that is amenable to analysis. There are several methods of transcribing conversational interaction and which one is used will depend on the analyst’s purpose. If, for example, the investigator is concerned with the minutiae of conversation such as which particular words are stressed, which words or utterances are spoken with increased loudness, the points at which the turns at talk overlap, how particular words and sounds are articulated, and so on, then a so-called narrow transcription may be required. If such detail is less important – perhaps because the analyst is more concerned with the overall gross structure of the conversation or the relative distribution of turns at talk amongst the participants – then a broad transcription may be sufficient. The chief distinction between these two approaches, therefore, is the fineness of the detail that is transcribed.
We can understand this better by looking at a few examples. We will use an extract from a longer conversation between a mother (M) and her teenage daughter (K) to exemplify both a narrow and a broad transcription. This will, therefore, demonstrate one way in which the same audio-recorded data can be represented differently.
Narrow transcription
Transcription conventions
In the same way that the investigator’s purpose will influence the choice of whether or not a broad or narrow transcription is required, so different researchers will highlight different detailed features of talk as being important to their purpose. Consequently, there are also several published transcription conventions. Having said this, I have found the following conventions to be particularly useful. They can almost certainly be profitably used for a number of purposes. They are essentially those outlined by Levinson (1983:369-370), with some minor modifications (Williamson, 1995).
Main Conventions: |
|
(.) |
micropause (comparable perhaps to an average syllable duration) <0.5 sec |
(..) |
brief pause >0.5 s <1.0 s |
(…) |
pause >1.0 s <1.5 s |
(2.0) |
longer pause in seconds |
// |
point at which the current utterance is overlapped by that transcribed below |
* |
asterisks indicate the alignment of the points where overlap ceases |
CAPS |
relatively high volume |
((CAPS)) |
analytical labels |
:: |
lengthened syllables or speech sounds |
– |
glottal stop, self editing marker |
= = |
latched utterances, with no gap |
? |
not a punctuation mark but a rising intonation contour |
(( )) |
used to indicate some phenomenon that the transcriber does not want to struggle with or some non vocal action |
( ) |
uncertain passages of transcript |
à |
draws attention to location of phenomenon of direct interest to discussion |
hh |
indicates an audible out-breath |
.hh |
indicates an audible in breath |
Other Conventions: |
|
Capital letters are not used except for proper nouns (e.g. Sunday, Graham), the 1st person pronoun, I (e.g. where should I go), and for indicating relatively high volume and analytical labels as set out above. |
|
Punctuation marks are not used, with the exception of the apostrophe (e.g. Baljeet’s; you’ve; they’ll). |
|
Line numbers in transcription extracts are referred to using the notation Lx, where L stands for ‘line’ and x is the relevant number. Thus, L01 refers to line number one and L25 refers to line number twenty-five, and so on. |
Example narrow transcription
The following example makes use of these conventions in setting out a narrow transcription.
01 M: o:ka:y (..) what do you want to talk about
02 K: I: don’t know
03 M: .hh
04 K: pick a conversation (..)
05 M: .hh w:: (.) what do you fan::c:y doing on s:aturday
06 K: Saturday (.) I (thought) we were going to get the
07 (.) the presents
08 M: yea:h (…) well w- we (.) talked about Saturday or Sunday
09 (…) er:: (..) bu:t Bede an::d (..) Sinners are
10 playing on Sunday so: (…) .hh wouldn’t give us much
11 ti:me to get back for four o clock (.) especially if we
12 wanted to: (.) go to Browton
13 K: m
14 M: so (..) probably Saturday
15 K: okay
16 M: er:m (…) but we’ve got the: (…)
17 K: ((sniff))
18 M: er::m (…) Paul Norton (.) and his wife coming round on
19 the (.) evening time (…) for a meal
20 K: oh (…) but (.) the only problem is I need to get (..)
21 dad’s present ((1 syllable))
22 M: mhm mhm
23 K: an:d we er:: (..) so: I can either do that on the Saturday
24 and Sunday but (…) I think (…) one of the shops might
25 not be open (…)
26 M: on Sunday (…) you you’re home all we- all=
27 K: =oh yeah
28 M: from Monday
29 K: oh will do yeah (…)
30 M: have you go:t er:m: (..) any plans
31 K: Friday (at home)
32 M: m
33 K: no: (…) I’ve got no idea what I’m doing (.) next week
34 (…)
35 M: so: (.) presumably though you’ll be going shopping (..)
36 K: OH YEAH (.) at some point (…)
37 M: how many presents have you got to buy yet (..)
38 K: I’ve just got to get (…) dad’s (…)
39 M: yeah
40 K: and I need to get (.) this thing for Emma (..) an:d (.)
41 see if I can see anything for Robert and (.) I’ll just
42 carry on with the present that I’ve got him (..) can you
Narrow transcriptions such as this are typically used by analysts using a technique known as conversation analysis (CA). Much of the work examining overlapping talk (interruptions) was done using this method.
Broad transcription
Notes
The narrow transcription above was used to generate the broad transcription presented below.
- Self-editing markers are glossed, e.g. “we-” (L26) is glossed as week.
- All non-linguistic phenomena, such as audible in-breaths, silences and laughter, have been removed, e.g. ((sniff)) in L17 has been omitted.
- All filled pauses have been removed, e.g. “have you go:t er:m: (..) any plans” (L30) is represented as have you got any plans.
- Successive repetitions have been removed, e.g. “w- we” (L08) is transcribed simply as we.
Example broad transcription
1 M: okay what do you want to talk about
2 K: I don’t know…pick a conversation
3 M: what do you fancy doing on Saturday
4 K: Saturday…I thought we were going to get the presents
5 M: yeah well we talked about Saturday or Sunday but Bede and
6 Sinners are playing on Sunday so wouldn’t give us much time to
7 get back for four o clock especially if we wanted to go to
8 Browton
9 K: m
10 M: so probably Saturday
11 K okay
12 M: but we’ve got the…Paul Norton and his wife coming round on
13 the evening time for a meal
14 K: oh but the only problem is I need to get dad’s present
15 ((1 syllable))
16 M: mhm
17 K: and we…so I can either do that on the Saturday and Sunday
18 but I think one of the shops might not be open
19 M: on Sunday…you, you’re home all week…all…
20 K: oh yeah
21 M: from Monday
22 K: oh, will do, yeah
23 M: have you got any plans
25 K: Friday, at home
26 M: m
27 K: no, I’ve got no idea what I’m doing next week
28 M: so presumably though you’ll be going shopping
29 K: oh yeah, at some point
30 M: how many presents have you got to buy yet
31 K: I’ve just got to get dad’s
32 M: yeah
33 K: and I need to get this thing for Emma and see if I can see
34 anything for Robert and I’ll just carry on with the present
35 that I’ve got him…can you think of anyone that I could get
36 him
It is evident that this broad transcription lacks the fineness of detail of the narrow transcription. However, it is still a useful representation of the talk between the two participants. For example, if we are mainly concerned with which particular word classes (verbs, nouns, adjectives, adverbs, numerals, determiners, pronouns, prepositions, conjunctions) each participant uses then the level of detail available in this broad transcription is sufficient for this purpose. It does not require an understanding of, for example, vowel length to determine word class and so the broad transcription adequately suits our research purpose.
Distribution of turns
The linear layout of both transcriptions shown above is a common arrangement. However, the distribution of turns at talk (i.e. the relative amount of speaker participation) is not immediately apparent in this format. The layout below highlights each participant’s contributions by citing them in a columnar format.
Example transcription showing turn distribution
M |
K |
||
1 |
okay what do you want to talk |
||
2 |
about |
||
3 |
I don’t know…pick a |
||
4 |
conversation |
||
5 |
what do you fancy doing on |
||
6 |
Saturday |
||
7 |
Saturday…I thought we were |
||
8 |
going to get the presents |
||
9 |
yeah well we talked about |
||
10 |
Saturday or Sunday but Bede and |
||
11 |
Sinners are playing on Sunday so |
||
12 |
wouldn’t give us much time to get |
||
13 |
back for four o clock especially |
||
14 |
if we wanted to go to Browton |
||
15 |
m |
||
16 |
so probably Saturday |
||
17 |
okay |
||
18 |
but we’ve got the…Paul Norton |
||
19 |
and his wife coming round on the |
||
20 |
evening time for a meal |
||
21 |
oh but the only problem is I need |
||
22 |
to get dad’s present |
||
23 |
((1 syllable)) |
||
24 |
mhm |
||
25 |
and we…so I can either do that |
||
26 |
on the Saturday and Sunday but I |
||
27 |
think one of the shops might not |
||
28 |
be open |
||
29 |
on Sunday…you, you’re home all |
||
30 |
week…all… |
||
31 |
oh yeah |
||
32 |
from Monday |
||
33 |
oh, will do, yeah |
||
34 |
have you got any plans |
||
35 |
Friday, at home |
||
36 |
m |
||
37 |
no, I’ve got no idea what I’m |
||
38 |
doing next week |
||
39 |
so presumably though you’ll be |
||
40 |
going shopping |
||
41 |
oh yeah, at some point |
||
42 |
how many presents have you got to |
||
43 |
buy yet |
||
44 |
I’ve just got to get dad’s |
||
45 |
yeah |
||
46 |
and I need to get this thing for |
||
47 |
Emma and see if I can see |
||
48 |
anything for Robert and I’ll just |
||
49 |
carry on with the present that |
||
50 |
I’ve got him…can you think of |
||
51 |
anyone that I could get him |
Whilst we have used our broad transcription extract to exemplify this layout, this design is not restricted to presenting just broad transcriptions. Narrow transcriptions could be similarly set out. However, this is less common. The layout works well when there are only two people involved in the conversation (so-called dyadic conversation) and can perhaps be extended to conversations with at least three participants. Anymore than this and you will appreciate that it becomes difficult to fit sufficient columns (of transcribed talk) across a printed page. However, this format can be usefully exploited using a computer spreadsheet where the number of columns is virtually limitless.
References
Levinson, S.C. (1983) Pragmatics Cambridge: Cambridge University Press.
Williamson, G. (1995) Instructor-trainee conversation in an Adult Training Centre for people with learning disabilities: an analysis of the function and distribution of back channel tokens and personal names Unpublished PhD thesis. Newcastle, UK: University of Newcastle.