In recent years there has been a growing interest in the sociological
study of environmental issues. One area in which this is evident is in
the application of social scientific methods in social impact assessm
ent (SIA): the study of the anticipated social impact of proposed chan
ges to the environment. This paper addresses one aspect of the debate
about appropriate methods for SIA; whether, and how, to include the ex
pressed views and perceptions of those who will be affected. It is arg
ued first that although SIA ostensibly deals with the social effects o
f projects there is a tendency for assessments to avoid any detailed c
onsideration of the ways in which people are affected. Instead there i
s an emphasis upon technical and economic considerations. When assessm
ents do attempt to incorporate the perceptions of local people they ty
pically do so through some form of attitude research. However if langu
age is viewed as a form of social action rather than as a detached com
mentary on reality there are radical implications for the methods trad
itionally used in SIA to gauge people's views and attitudes. I conclud
e by outlining an alternative to the traditional practice of viewing p
eoples' accounts as a repository of their attitudes, arguing that they
might be more profitably used to explore how social impacts are socia
lly constructed.