Method I 97
evidence of acquisition, or the 'risk of making the opposite, false
negative error - that of treating non-occurrence as evidence of non-
acquisition, (Veils,1985:132). Some of these errors, I feel now, were in
fact made in the IMDE study presented in 3.4. The present study attemps
to avoid these pitfalls in that it does not pretend to be a pronouncement
on acquisition, only on frequency of use.
Vithin the naturalistic paradigm, a number of different recording
techniques are available. Reynolds (1986) discusses the advantages and
disadvantage of four of them in a valuable paper on Participant
Observation with Children in Southern Africa, and as a result of years
of anthropological work she appeals for 'consciousness of the limitations
inherent in any recording technique' (p. 25) and suggests the use of
combinations of them . The discussion of different recording techniques
Is superfluous here, given that my study is an extension of a previous
investigation (the IMDE Project), and it is based on a subset of data
selected from the larger corpus of data collected for that research. The
method and techniques of data collection are described in 3.4 and in the
following pages.
Veils similarly concludes that the combination of naturalistic and
experimental methods, ideally in a cyclical relationship within the same
research, would be the best solution, for they are complementary. Indeed,
this was the original design of the INDE project: semi-experimental
situations were created by means of games and other set activities and
the elicited talk recorded. But in our small IMDE team, as in the Bristol
large one, 'we did not succeed in articulating the two approaches as we
Intended, for the major part of our resources were taken up by the more
labour-intensive collection and analysis of the samples of spontaneous
conversation' (Veils,1985:128). As for my study, complementary methods of
data collection, such as through elicitation under controlled conditions,
or further checks using triangulation (Adelman,1981; Cohen and
Manion,1985) had to be ruled out because of my insufficient competence in
the subjects' LI.